EFFECTS OF ALTERNATIVE GRASS SPECIES ON GRAZING
PREFERENCE OF SHEEP FOR WHITE CLOVER
A thesis submitted in partial fulfilment of the requirements for the Degree of
Master of Agricultural Science
at
Lincoln University
Canterbury
New Zealand
by
Tomohiro Muraki
2008
ABSTRACT
Abstract of a thesis submitted in partial fulfilment of the requirements for the Degree of
Master of Agricultural Science
EFFECTS OF ALTERNATIVE GRASS SPECIES ON GRAZING PREFERENCE
OF SHEEP FOR WHITE CLOVER
by
Tomohiro Muraki
Despite the importance of a high white clover (Trifolium repens) content in temperate
pastoral systems in terms of livestock performance and nitrogen fixation, the proportion of
white clover in grass-clover pastures is often low (<20%). This thesis examined in two
experiments whether the white clover content of pastures could be improved by sowing
white clover with alternative grass species to diploid perennial ryegrass (Lolium perenne
L.).
In a pasture experiment, DM production, pasture composition and morphology of
grass-clover mixtures was measured over the establishment year (January 2007 to January
2008) where white clover was sown in fine mixtures with diploid perennial ryegrass,
tetraploid perennial ryegrass, timothy (Phleum pratense L.) and cocksfoot (Dactylis
glomerata L.). Pastures were irrigated and rotationally grazed with on-off grazing with
Coopworth ewe hoggets. Total annual DM production of pasture was more than 20%
higher in tetraploid (12521 kg DM ha-1) and diploid (11733 kg DM ha-1) perennial ryegrass
than timothy (9751 kg DM ha-1) and cocksfoot (9654 kg DM ha-1). However, timothy
(5936 kg DM ha-1) and cocksfoot (5311 kg DM ha-1) had more than four times higher
white clover annual DM production than tetraploid (1310 kg DM ha-1) and diploid (818 kg
DM ha-1) ryegrass. Pasture growth rate at the first three harvests in autumn was
significantly greater in tetraploid and diploid ryegrass than timothy and cocksfoot. Timothy
and cocksfoot had a higher proportion of white clover than tetraploid and diploid perennial
ryegrass throughout the entire year. This was due to more and larger white clover plants in
timothy and cocksfoot plots.
In a grazing preference experiment, the partial preference of sheep for white clover offered
in combination with the same grass species as in the pasture experiment was measured in
ii
five grazing tests in May, September, October, November and December 2007. Pastures
were sown in January 2007. Paired plots (grass and clover both 4.2 m x 10 m) were grazed
by three Coopworth ewe hoggets between 9am and 5pm, and preference was recorded by
decline in pasture mass and visual scan sampling for grazing time. Grazing preference for
clover was generally low throughout these tests (e.g. average apparent DM intake from
clover = 47%; average grazing time from clover = 44%). Several explanations are
proposed for this low preference including a high N content and intake rate of the grass
relative to the clover. No significant differences were found among the grass treatments in
total grass grazing time, total clover grazing time, ruminating time, the proportion of
grazing time on clover, selective coefficient for clover and DM intake percentage from
clover at any date. There was no significant change in overall sward surface height (SSH)
decline among grass treatments throughout all the tests except December 2007 when the
overall SSH decline for cocksfoot was significantly lower than the other species.
The study indicated that the rapid growth rate of perennial ryegrass in the early phase of
pasture establishment, rather than differences in partial preference, was the key factor
limiting white clover content in the mixed swards relative to cocksfoot and timothy
pastures. It is concluded that high clover-containing pastures capable of delivering high per
head performance can be established through the use of slow establishing pasture species
such as timothy and cocksfoot.
KEYWORDS: tetraploid perennial ryegrass, diploid perennial ryegrass, timothy, cocksfoot,
white clover, Lolium perenne, Phleum pratense, Dactylis glomerata, Trifolium repens,
grazing preference, mixed swards, monocultures, growth rate.
iii
TABLE OF CONTENTS
ABSTRACT………………………………………………………………………………...ii
TABLE OF CONTENTS…………………………………………………………………..iv
LIST OF TABLES…………………………………………………………………………vii
LIST OF FIGURES………………………………………………………………………...ix
LIST OF PLATES…………………………………………………………………………..x
1
CHAPTER ONE: GENERAL INTRODUCTION………………………………. 1
1.1 Background………………………………………………………………………. 1
1.2
2
2.1
2.2
Objectives………………………………………………………………………… 3
CHAPTER TWO: LITERETUR REVIEW……………………………………….4
Scope of literature review…………………………………………………………4
Significance of white clover………………………………………………………4
2.2.1
2.2.2
2.2.3
Nitrogen fixation……………………………………………………………..4
Nutritive and feeding value…………………………………………………..5
Seasonal growth pattern……………………………………………………...9
2.2.4
Background of low clover content…………………………………………...9
2.3 Effects of diet selection…………………………………………………………. 10
2.3.1
Diet selection………………………………………………………………..10
2.3.2
Partial preference……………………………………………………………10
2.3.3
Other potential elements regulating grazing preference…………………….11
2.4 Increasing clover content of temperate pastures…………………………………13
2.4.1
Clover breeding and sowing………………………………………………...13
2.4.2
Spatial separation of monocultures…………………………………………13
2.4.3
N fixation and uptake……………………………………………………….15
2.4.4
Alternative grass species……………………………………………………15
2.5 Compatibility of grass pasture species with white clover in mixtures…………...17
2.6 Aim……………………………………………………………………………….18
2.7 Specific objectives of this study …….…………………………………………...18
3 CHAPTER THREE: THE EFFECT OF GRASS SPECIES ON PASTURE
PRODUCT ION, C OMPOS IT ION, SEEDLING ESTABLIS HMENT AND
MORPHOLOGICAL PROPERTY OF MIXED GRASS-CLOVER PASTURES……..19
3.1 Materials and methods……………………………………………………………19
3.1.1
Experimental site and design………………………………………………..19
3.1.2
Pasture establishment……………………………………………………….20
3.1.2.1 Irrigation…………………………………………………………………..21
3.1.2.2 Grazing……………………………………………………………………21
3.1.3
Measurements……………………………………………………………….23
3.1.3.1 DM production and composition………………………………………….23
3.1.3.2
Grass and white clover populations……………………………………..24
iv
3.1.3.3
Grass and clover plant morphology……………………………………..24
3.1.3.4
Pasture nutritive value…………………………………………………..25
3.1.4
Calculations…………………………………………………………………25
3.1.4.1 Total DM yield of pasture, grass, clover, weeds and dead material from
sowing to the end of grazing trials………………………………………...25
3.1.4.2 Daily growth rates of pasture, grass, clover and weeds at each grazing
time………………………………………………………………………..25
3.1.4.3 Botanical composition…………………………………………………….26
3.1.5
Weather data………………………………………………………………...26
3.1.6
Statistical analysis…………………………………………………………..26
3.2 Results……………………………………………………………………………27
3.2.1
Climate……………………………………………………………………...27
3.2.2
3.2.3
3.2.4
Total DM yield……………………………………………………………...29
Pasture growth rates………………………………………………………...30
Botanical composition………………………………………………………31
3.2.5
Plant density………………………………………………………………...34
3.2.6
Clover and grass morphology………………………………………………35
3.2.7
Pasture nutritive value………………………………………………………38
3.3 Discussion……………………………………………………………………......38
3.3.1
3.3.2
3.3.3
White clover yield: timothy and cocksfoot versus ryegrass species………..38
White clover content of diploid versus tetraploid perennial ryegrass………41
Total Pasture Production……………………………………………………42
4 CHAPTER FOUR: THE EFFECT OF GRASS SPECIES ON GRAZING
PREFERENCE OF SHEEP FOR WHITE CLOVER AND GRASS…………………...48
4.1 Materials and methods……………………………………………………………48
4.1.1
Experimental site and design………………………………………………..48
4.1.2
Pasture management………………………………………………………...49
4.1.2.1
Cultivars and sowing rate……………………………………………….49
4.1.2.2
Irrigation………………………………………………………………...50
4.1.3
Preference tests………………………………………………………….......50
4.1.4
Measurements……………………………………………………………….52
4.1.4.1
Animal behaviour measurements……………………………………….52
4.1.4.2
Pasture measurements…………………………………………………..54
4.2 RESULTS………………………………………………………………………...55
4.2.1
May preference test…………………………………………………………55
4.2.1.1
Pasture characteristics…………………………………………………..55
4.2.1.2
Grazing behaviour and pasture changes………………………………..58
4.2.2
September preference test
………………………………………………61
4.2.2.1
Pasture characteristics…………………………………………………..61
v
4.2.2.2
Grazing behaviour and pasture changes………………………………...62
4.2.3
October preference test……………………………………………………...64
4.2.3.1
Pasture characteristics…………………………………………………..64
4.2.3.2
Grazing behaviour and pasture changes………………………………...67
4.2.4
November preference test…………………………………………………...69
4.2.4.1
Pasture characteristics…………………………………………………..69
4.2.4.2
Grazing behaviour and pasture changes………………………………...72
4.2.5
December preference test…………………………………………………...73
4.2.5.1
Pasture characteristics…………………………………………………..73
4.2.5.2
Grazing behaviour and pasture changes………………………………...76
4.2.6
Summary of percentage of grazing time on white clover………………..…78
4.3 DISCUSSION……………………………………………………………………80
4.3.1
Low partial preference………………………………………………………80
4.3.1.1
Preference for grasses…………………………………………………...81
4.3.1.2
N content………………………………………………………………..83
4.3.1.3
Not getting complete preference test……………………………………83
4.3.1.4
Novelty………………………………………………………………….84
4.3.1.5
Fibre and N concentration………………………………………………85
4.3.2
No grass species effect on preference………………………………………86
4.3.2.1
4.3.2.2
4.3.2.3
Timothy…………………………………………………………………86
Cocksfoot……………………………………………………………….87
Effect of pasture mass and height………………………………………87
5
6
4.3.2.4
Effect of WSC on grazing preference……………………………….….91
4.3.3
Possible nutritive value concept…………………………………………….92
CHAPTER FIVE : GENERAL DISCUSSION…………………………………..94
CHAPTER SIX : CONCLUSIONS…………………………………………….101
7
8
REFERENCES………………………………………………………………….103
ACKNOWLEDGEMENTS…………………………………………………….112
vi
Table 2.1
Table 2.2
Table 2.3
Table 3.1
Table 3.2
Table 3.3
Table 3.4
LIST OF TABLES
The feeding values of several pasture species in terms of sheep live
weight...............................................................................................................7
The comparative feeding values of pastures in dry matter intake (DMI) and
milksolids (MS) production of cows during a mid-late-lactation period.........7
Chemical composition (%DM) of perennial ryegrass and white clover in New
Zealand……………………………………………………………………….8
Soil pH, Olsen P and calcium, magnesium, sodium and sulphur (all MAF
quick the units) for experimental site on 20 January 2007…………………19
Pasture species, cultivars and sowing rates used in the DM production,
composition and morphology trial………………………………………….20
Monthly rainfall at Lincoln University from January 2007 to January
2008.………………………………………………………………………...28
Average monthly air temperature at Lincoln University from January 2007 to
January 2008.………………………………………………………………..29
Table 3.5
Total dry matter production of pasture, grass, clover, weed and dead material
components (kg DM ha-1) from 8 January to 6 December 2007.…………..30
Table 3.6
Number of plants per ㎡ in mixed swards and pure monoculture swards in
autumn (March) 2007 and in summer (January) 2008.……………………..35
Table 3.7
The number of tillers per plant, tallest shoot length, and root and shoot mass
of individual grass species in March 2007 and in January 2008.…………..36
The number of stolons, fully opened leaves, length of longest above-ground
Table 3.8
Table 3.9
Table 3.10
Table 4.1
Table 4.2
Table 4.3
Table 4.4
Table 4.5
Table 4.6
part, and root and shoot weight per plant of white clover plants in March
2007 and January 2008……………………………………………………...37
Concentrations of metabolisable energy (MJ ME kg-1 DM) of grass and white
clover in the binary mixtures on 7 November 2007………………………..38
Total dry matter of overall pasture after each grazing occasion (kg DM ha-1)
from 8 January to 6 December 2007………………………………………40
Pasture species, cultivars and sowing rates used in the grazing preference
trial.………………………………………………………………………..50
Pasture mass, composition and morphology of grass and clover plots in each
grass treatment combination prior to the May preference test.……………56
Nutritive value of grass and clover determined by NIRS prior to the May
preference test..………………………………………….............................57
The concentration of macronutrients (%DM) contained in laminae of grass
and clover prior to the May grazing trial………………………………….58
Grazing behaviour and changes in pasture for the May preference test.…..59
Pasture mass and pasture composition of grass and clover plots in each grass
treatment combination prior to the September grazing……………………61
vii
Table 4.7
Table 4.8
Table 4.9
Table 4.10
Table 4.11
Table 4.12
Table 4.13
Table 4.14
Table 4.15
Table 4.16
Table 4.17
Table 4.18
Table 4.19
Table 4.20
Table 5.1
Nutritive value of grass and clover determined by NIRS prior to the
September preference test.………………………………………………...62
Grazing behaviour and changes in pasture for the September preference
test.……………………………………………………………..…………63
Pasture mass and pasture composition of grass and clover plots in each grass
treatment combination prior to the October grazing………………………65
Nutritive value of grass and clover determined by NIRS prior to the October
preference test.……………………………………………………………66
The percentage of macronutrients (%DM) contained in laminae of grass and
clover prior to the October grazing……………………………………….67
Grazing behaviour and changes in pasture for the October preference
test.…………………………………………………………….………....68
Pasture mass and pasture composition of grass and clover plots in each grass
treatment combination prior to the November grazing…………………...70
Nutritive value of grass and clover determined by NIRS prior to the
November preference test…………………………………………………71
Grazing behaviour and changes in pasture for the November preference
test………………………………...……………………………………….72
Pasture mass and pasture composition of grass and clover plots in each grass
treatment combination prior to the December grazing……………………74
Nutritive value of grass and clover determined by NIRS prior to the
December preference test…………………………………………………75
The percentage of macronutrients (%DM) contained in laminae of grass and
clover prior to the December grazing……………………………………76
Grazing behaviour and changes in pasture for the December preference
test.………………………………………………………...………………77
Correlation between average percentage of grazing time on clover plots and
NDF% in grass species, between clover grazing time% and N% in grass,
between the mean proportion of grazing time on clover and a ratio of NDF
to N at each test from May to December 2007……………………………80
Annual dry matter production of grass, clover and the sum of grass and
clover in pure swards from 8 January to 6 December 2007 and mixed
swards from 8 January to 14 December 2007 (kg DM ha-1)………………98
viii
Figure 3.1
Figure 3.2
Figure 3.3
Figure 4.1
Figure 4.2
Figure 4.3
Figure 4.4
Figure 4.5
Figure 4.6
Figure 4.7
Figure 4.8
LIST OF FIGURES
Plot design for the pasture production, composition, seedling establishment
and morphological study.…………………………………………………20
Pasture growth rates of grass-white clover pastures from 8 January 2007 to 6
December 2007.……………………………………………………………31
The botanical composition (% by DM in pre-grazing mass) of (a) grass, (b)
white clover, (c) weeds (dicot and grass) and (d) dead material in the four
grass-white clover mixtures from March to December 2007.……………..33
Plot design for grazing preference trial……………………………………49
Sward surface height (cm) of (a) grass and (b) clover plots in paired
comparisons during the May trial.………………......................................60
Sward surface height (cm) of (a) grass and (b) clover plots in paired
comparisons during the September trial.…………………………………64
Sward surface height (cm) of (a) grass and (b) clover plots in paired
comparisons during the October trial……………………………………69
Sward surface height (cm) of (a) grass and (b) clover plots in paired
comparisons during the November trial.…………………………………73
Sward surface height (cm) of (a) grass and (b) clover plots in paired
comparisons during the December trial.…………………………………78
Proportion of grazing time on clover plots at each test from May to
December 2007.………………………………………………………..…79
Comparison of neutral detergent fibre concentration of dry matter between
(a) snipped grasses (tetraploid and diploid perennial ryegrass, timothy and
cocksfoot) prior to grazing, and (b) possible nutritive values (pNVs) (%)
calculated from the percentage of grazing time on grass and clover by
treatments at each date.…………………………………………………..93
ix
Plate 3.1
Plate 3.2
Plate 3.3
Plate 3.4
Plate 3.5
Plate 3.6
Plate 3.7
Plate 3.8
Plate 4.1
Plate 4.2
LIST OF PLATES
Photo showing experimental plots in August 2007…………………………..22
Photo showing experimental plots during grazing in August 2007…………..22
Photo showing a diploid perennial ryegrass plot after grazing in August
2007……………………………………………………………………..…..23
Seedlings in a tetraploid perennial ryegrass / white clover sward 45 days after
sowing (a and b). (22 February 2007, 10 days before 1st plant number
measurement)……………………………………………………..………...43
Seedlings in a diploid perennial ryegrass / white clover sward 45 days after
sowing (a and b). (22 February 2007, 10 days before 1st plant number
measurement)…………………………………………………..…………..44
Seedlings in a timothy / white clover sward 45 days after sowing (a and b).
(22 February 2007, 10 days before 1st plant number measurement)………45
Seedlings in a cocksfoot / white clover sward 45 days after sowing (a and b)
(22 February 2007, 10 days before 1st plant number measurement).………46
Tetraploid ryegrass / white clover mixture (above, a) and timothy / white
clover mixture (below, b) on 20 Jun 2007, six months after sowing……….47
View of a preference test site from the observation tower (September 2007).
………………………………………………………………………………51
Plate 4.3
View of a preference test site from the observation tower (September
2007)………………………………………………………………………...53
A sheep in idle state…………………………………………………………53
Plate 4.4
Plate 4.5
Plate 4.6
Plate 4.7
Fully grazed cocksfoot (October 2007)……………………………………..82
Fully grazed tetraploid perennial ryegrass (September 2007)………………82
Hardly grazed a tall clover patch after the preference test (October 2007)….89
Fully grazed clover with low sward height (October 2007)………………...89
Plate 4.8
Ungrazed tall clover in the same paddock as Plate 4.7 with a different angle
(October 2007)……………………………………………………………...90
Contrast between fully grazed and ungrazed clover after preference test
(December 2007)……………………………………………………………91
Plate 4.9
x
1 CHAPTER ONE
GENERAL INTRODUCTION
1.1 Background
Pastures containing a mixture of perennial ryegrass (Lolium perenne) and white clover
(Trifolium repens) form the basis of low-cost animal production systems in temperate
regions of the world (Caradus et al., 1996). White clover provides nitrogen (N) to the
pasture via nitrogen fixation, and is also a forage of high feeding value relative to grass.
Numerous studies show that as the proportion of white clover in the pasture increases there
is improved performance of both sheep, beef and dairy cows (Caradus et al., 1996; Parsons
et al., 2006), with performance on pure white clover swards generally exceeding
performance on mixed grass-clover pastures (Cosgrove et al., 2003). Thus, the ability to
sustain a high proportion of clover in the pasture is important for N input into pastures and
high animal performance.
Ideally at least 25-45% of the total herbage grown dry matter (DM) in grass-white clover
pastures should be clover to achieve adequate N inputs and high animal performance
(Thomas, 1992; Kemp et al., 1999). However, in most temperate pastures the average
white clover content over the year is less than 20% (Ettema and Ledgard, 1992). Thus,
white clover contents are usually too low to support the full annual DM production
potential of ryegrass-white clover pastures. This leads to greater reliance on nitrogenous
fertilisers, which in turn may reduce legume content further particularly if stocking rate is
not increased to harvest the extra feed resulting from N fertilizer (Clark and Harris, 1996).
With increases in the price of N fertilizer and concerns over the impacts of excessive N
1
fertiliser usage causing increased N losses via leaching and to atmosphere, there is clearly
a need to develop methods to improve the proportion of clover in pastures.
One strategy to improve the legume content of pastures has been to develop via plant
breeding improved white clover cultivars. However, although new cultivars have
performed well in small plot trials, their effect on clover content and animal performance
when evaluated at the farm scale has been relatively small (Crush et al., 2006). Parsons et
al. (2006) reviewed a range of other strategies that may be used to increase the white
clover content of pastures. These included the use of pure white clover and grass swards as
spatially separated monocultures, the manipulation of N fixation and N uptake
relationships between grasses and clovers, and the alteration of competition and diet
selection through the use of alternative grass species to perennial ryegrass. Parsons et al.
(2006) noted that to develop practical strategies to increase the clover content of the
pasture and diet requires an understanding of animal behaviour, including grazing
preference and diet selection. This is because one of the key factors affecting grass-legume
interactions is diet selection and the extent to which clover and grass are grazed within a
pasture. Previous studies show that livestock prefer a diet of around 60-80% white clover
when white clover is offered in combination with perennial ryegrass (Newman et al.,
1994b; Penning et al., 1995b; Cosgrove et al., 1996; Torres-Rodriguez et al., 1997; Rutter
et al., 2005b). As mixed pastures rarely contain this amount of white clover, livestock often
select clover out of the pasture (Cosgrove and Edwards, 2007). In turn, this places clover at
a disadvantage in terms of plant to plant competition and contributes to low clover content
in the sward.
2
1.2 Objectives
This thesis focuses on the use of alternative grass species to diploid perennial ryegrasses as
a means to improve white clover content. The compatibility of several kinds of grass
species, － diploid and tetraploid perennial ryegrass, timothy (Phleum pratense) and
cocksfoot (Dactylis glomerata) － with white clover are explored. It is proposed that the
white clover content may be higher with alternative grass species as they are less
competitive with white clover and because they have grazing preference closer to that of
white clover (Edwards et al., 2008).
The objectives of this work were:
1. To determine the effect of grass species on pasture production, composition and
morphology of mixed grass-white clover pastures over the establishment year, and
2. To determine the effect of grass species on partial preference of sheep for white clover.
3
2 CHAPTER TWO
LITERATURE REVIEW
2.1 Scope of literature review
This literature review concentrates on grass-clover pastures for either irrigated or summer
moist environments. Brock (2006) indicated that the amount of rainfall in summer that was
needed for white clover survival was 40 mm/month and 60 mm/month for production. The
significance of white clover in New Zealand pastures, the background of low clover
content and how diet selection affects clover content is reviewed. Methods to improve the
white clover content are then discussed. Finally, aims and specific objectives for this study
are given.
2.2 Significance of white clover
2.2.1
Nitrogen fixation
Nitrogen (N) fixation is a remarkable characteristic which legumes possess, and its process
is carried out by rhizobia, the generic name of Azorhizobium, Bradyrhizobium,
Photorhizobium, Rhizobium and Sinorhizobium bacteria (McKenzie et al., 1999; Taiz and
Zeiger, 2002). These nitrogen-fixing bacteria convert atmospheric N, di-nitrogen gas (N2),
into ammonium (NH4+), which is easily available for plants, with the nitrogenase enzymes
(Taiz and Zeiger, 2002). Apart from other kind of free-living microorganisms in soils,
4
rhizobia have a symbiotic relationship with legume species, such as white clover. In return
for providing nitrogen, these bacteria obtain carbohydrates as an essential nutrient and
anaerobic environment within the root tissues of the host legume, particularly in nodules
(Langer, 1990; McKenzie et al., 1999; Taiz and Zeiger, 2002).
The contribution of N to pastures via N fixation is one of the primary grounds that legumes
are chosen as a pasture species. Nitrogen fixed by legume is passed to grasses when clover
dies or via dung and urine. Potential N-fixation rates from white clover range 600-700 kg
N ha-1 year-1 (Crush 1987). However, the presence of mineral N (such as from N fertilizer)
and factors which reduce white clover growth and abundance (e.g. grass competition, diet
selection) result in much lower N fixation rates. Thus, reported N fixation rates range from
17 kg N ha-1 year-1 in infertile hill pastures to 380 kg N ha-1 year-1 in intensively managed
pastures (Crush, 1987). In the absence of mineral N, there is a direct positive relationship
between N fixation and white clover growth (Hoglund and Brock, 1987). Therefore, from
the point of view of improving N inputs into pastures, it is important to increase the white
clover content of the pasture. Thomas (1992) estimated that legume contents of 20-45%
could provide the N requirements of a sustainable productive pasture. However, the
legume content of most pastures is less than 20%, and thus N supply is too low to support
the full productive potential of ryegrass based pastures (Parsons et al., 2006).
2.2.2
Nutritive and feeding value
White clover is regarded as a high quality forage for livestock in grazed temperate pastures.
High quality results from a greater nutritive and feeding value for white clover compared
with other pasture species such as perennial ryegrass (Lolium perenne L.). Nutritive value
refers to the animal response per unit of feed intake and reflects available nutrients which
5
are required by animals, including crude protein, lipids, fat-soluble vitamins,
macro-elements (sodium, calcium, potassium, phosphorus, sulphur, magnesium, chlorine),
micro-elements (cooper, cobalt, selenium, iodine, iron, zinc, manganese) and energy
(metabolisable or digestible energy) (Ulyatt, 1981; Waghorn and Clark, 2004). Feeding
value refers to animal production response (e.g. g head-1 day-1) and therefore is a function
of both intake and nutritive value (Ulyatt, 1981). Compared to grass species, including
perennial ryegrass, white clover contains higher crude protein, readily fermentable
carbohydrate (e.g. water-soluble sugars and pectin), and metabolisable energy (ME) but
lower structural carbohydrate (e.g. cellulose), lignin and fibre (Ulyatt et al., 1977; Waghorn
and Clark, 2004).
Relative feeding values for sheep and cattle are consistently higher in white clover than
grasses (Tables 2.1 and 2.2). The dietary advantages of white clover over grass species may
be attributed to a higher proportion of non-structural carbohydrates and lower content of
structural carbohydrates (Ulyatt et al., 1977). Compared with structural carbohydrates such
as hemicellulose and cellulose, non-structural carbohydrates, including water-soluble
sugars and pectin, are more fermentable, and more easily digested in the rumen (Ulyatt et
al., 1977). Hence, a higher ratio of non-structural carbohydrates to structural carbohydrate
is nutritionally favourable (Ulyatt et al., 1977). The proportions of readily fermentable
carbohydrate and structural carbohydrate, and ratios of these carbohydrates in ryegrass and
white clover are seen in Table 2.3. The considerably higher value of the ratio for white
clover (1.17) than that for ryegrass (0.42) indicates that ryegrass is more slowly digested in
the rumen, and in turn the nutritive content of clover is more likely to be utilised for animal
production.
In addition, feeding values of grass species can vary with the presence and degree of
6
endophyte infection. For instance, Edwards et al. (1993) demonstrated that the intake of
endophyte-infected (wild type) perennial ryegrass by sheep was lower than that of
endophyte-free grass within the same species, and suggested a subsequent reduction of
animal performance. Therefore, when endophyte-infected, the grass species may show
lower animal performance even if the feeding value (e.g. ME) of the grass was high.
Furthermore, the nutritive value of white clover declines less than that of grass species
with the onset of reproductive development, increasing temperatures and with long
regrowth intervals than perennial ryegrass does (Waghorn and Clark, 2004). Thus, legume
dominated pasture may be easier to manage for high nutritive value, than a grass
dominated pasture.
Table 2.1 The feeding values of several pasture species in terms of sheep live weight
(adopted from Ulyatt 1981).
Species
Perennial ryegrass `Grasslands Ruanui' (diploid)
Perennial ryegrass `Grasslands Ariki' (tetraploid)
Perennial ryegrass `Grasslands Manawa' (diploid)
Timothy, Common cultivar
White clover `Grasslands Huia'
Relative feeding value
100
111
148
129
192
Table 2.2 The comparative feeding values of pastures in dry matter intake (DMI) and
milksolids (MS) production of cows during a mid-late-lactation period (adopted from
Johnson and Thomson 1996).
Perennial ryegrass `Yatsyn-1' (diploid)
Timothy `Grasslands K ahn'
Cocksfoot `Grasslands Kara'
White clover `Grasslands Kopu'
Total nitrogen content Relative feeding value Relative feeding value
(%)
in DMI
in MS
4.02
100
100
4.24
104
109
4.03
117
109
4.26
96
168
7
Table 2.3 Chemical composition (%DM) of perennial ryegrass and white clover in New
Zealand (Ulyatt et al., 1977).
Cutting height
(cm)
Ryegrass (`Ruanui')
15
White clover (`Huia')
10
Readily
Structural
fermentable
carbohydrate
carbohydrate (a)
(b)
12.3
29.5
20.3
17.3
a /b
0.42
1.17
The intake rate of sheep and cattle (g DM min-1) increases as the proportion of white clover
in the pasture increases, with consistently greater intake rates in pure white clover than
ryegrass swards (Penning et al., 1991; Edwards et al., 1995; Penning et al., 1995b). This is
partly due to larger bite masses from white clover than ryegrass, although the difference in
intake rate may be smaller than expected due to an associated decline in prehension rate as
bite mass increases. The larger bite masses are in turn due to larger bite areas, with there
being little difference in bite depth between ryegrass and white clover (Edwards et al.,
1995). However, the factor contributing most to the higher intake rates of white clover than
ryegrass is the lower mastication cost for clover than for grass (Newman et al., 1992;
Parsons et al., 1994b; Penning et al., 1995b). In general, higher intake rates lead to greater
daily intakes in white clover than perennial ryegrass (Penning et al., 1995b). However, in
some cases, animals reduce grazing time, and increases in daily intake do not arise
(Penning et al., 1995b).
The greater nutritive value and higher intake rate combine to give improved feeding value
and livestock performance as the proportion of white clover in the pasture increases. For
example, in a dairy study, milk production per cow increased from 0.80 kg milk solids
(MS) cow-1 day-1 at 20% clover content to 0.93 kg MS cow-1 day-1 at 50% clover content in
the diet (Harris et al., 1997). In lamb growth studies, Gibb and Treacher (1984) found a
high correlation (R = 0.78) of empty-live-weight gain per head with the proportion of
8
clover as organic matter in a diet. In addition, growth of lambs is consistently higher on
pure clover pastures than grass-clover mixtures. For instance, average daily gain of weaned
lambs was much higher in clover monoculture swards (345 g head-1 day-1) than in the
mixtures of grass and clover (205 g head-1 day-1) (Cosgrove et al., 2003).
2.2.3
Seasonal growth pattern
A further useful feature of perennial ryegrass and white clover mixtures reflects their
complementary growth patterns. Ryegrass and clover have complementary patterns of
seasonal growth with grasses being more productive during autumn, winter and spring, and
clover potentially (when given adequate water) more productive during summer (Brock,
2006; Parsons et al., 2006).
2.2.4
Background of low clover content
Perennial ryegrass-white clover pastures in New Zealand frequently contain less white
clover than necessary to meet the requirement for high animal production (Chapter 1).
With regard to the causes of low clover content in a mixed pasture of New Zealand
pastoral farming, Brock (2006) mentioned three factors. First, seedling establishment
methods that enhance grass establishment, rather than clover establishment, such as high
sowing rates of grass, direct drilling at high speed, sowing clover too deep and sowing
clover too late in autumn. Second, management to maximise grass mass, such as increased
N fertiliser use, infrequent and lenient grazing and selection of more productive grasses.
Third, selective grazing for clover over grasses.
9
2.3 Effects of diet selection
Diet selection is where the diet harvested differs from the diet on offer in the sward and
reflects some sward components but not others being harvested. For example, more clover
is generally selected than that which is on offer (Newman et al., 1992; Parsons et al.,
1994a; Penning et al., 1995a; Cosgrove et al., 1996). This places clover at a disadvantage
in terms of plant to plant competition with ryegrass (reviewed in Cosgrove and Edwards
2007), particularly when ryegrass is dense and tall. Although selecting clover may increase
the proportion of clover in diet initially, it may limit clover growth and make clover less
competitive with grass, and finally may reduce the amount available in a mixed pasture.
2.3.1
Diet selection
Diet selection is a function of grazing preference modified by environmental constraints.
Thus, as a first step in determining the impact of grazing livestock on pasture composition
it is important to have knowledge of the grazing preference of livestock.
2.3.2
Partial preference
Considerable research now shows that sheep, cattle and goats all exhibit a partial
preference for white clover over perennial ryegrass. When offered freely voluntary choices
between pure swards of white clover and perennial ryegrass, they select 60-80% white
clover as a diet, but always continue to eat perennial ryegrass to some extent (Newman et
al., 1994a; Parsons et al., 1994a; Penning et al., 1995a; Cosgrove et al., 1996). This takes
place despite the situation where these animals are able to acquire a diet of pure clover at
no extra foraging cost. There is also a diurnal pattern of preference, with sheep and cattle
preferring clover in the morning, but eating more ryegrass as the day passes (Parsons et al.,
10
1994a; Penning et al., 1995a).
It has not been clearly deciphered yet why partial grazing preference would take place
(Rutter, 2006). Soder et al. (2007) reviewed the basis of this partial preference, listing a
range of hypothesis. These include maintenance of rumen function (Rutter et al., 2000),
maintenance of appropriate C:N ratio in rumen (Dewhurst et al., 2000; Merry et al., 2002;
Rutter, 2006), predation hazard and the perceived risk of predators (Newman et al., 1995)
and avoidance of toxins (Kyriazakis and Oldham, 1993). Rutter (2006) concluded that
ruminant animals possibly select a diet to acquire a balance between dual goals, which
might compete between them, and it is not necessarily that partial preference occurred due
to a single common basis supporting these goals. In this context, it is important to note that
the grazing preference results mentioned above are based on white clover in comparison to
diploid perennial ryegrass cultivars. It is not known whether the partial preference of
60-80% would be different for other grass species if tetraploid versus diploid perennial
ryegrasses were used.
2.3.3
Other potential elements regulating grazing preference
Numerous nutritive and mineral constituents of a pastoral diet have been implicated in
grazing preference. The digestibility and ME of a pasture species may be one of the factors
for grazing preference since the diet selectively grazed generally contain higher
digestibility and ME than those of the average pastures in the same paddock (Jamieson and
Hodgson, 1979). This can be happened under lax grazing to a larger extent, and is less
likely to be exhibited in intensive grazing systems as the opportunities for selecting a
specific diet are greater in the former case and limited in the latter one (Cosgrove and
11
Edwards, 2007). Mineral contents (e.g. calcium and potassium content and/or various
mineral ratios, etc) of a pasture plant show both positive and negative correlation with
palatability amongst several cultivars of cocksfoot offered to heifers (Mizuno et al., 1998).
Edwards et al. (1993) suggested that the negative grazing preference of sheep for cocksfoot
might have occurred due to the low N content of cocksfoot herbage. The relatively higher
degree of grazing preference of sheep has also been seen among several weeds with
relatively high mineral content in a tussock grasslands (Hughes, 1975). Hence,
measurement of these chemical constituents is necessary in this study.
Besides nutritive aspects, morphological features of a pasture species have been thought as
one of the possible factors which affect animal intake behaviour. Orr et al. (2004)
suggested that ingestive characteristics of heifers might be influenced by sheath tube and
leaf length when grass was grazed and by herbage mass as white clover was ingested. The
texture of a plant is possibly a limiting factor for grazing preference in a certain situation,
especially, where animals can select the favourable diet from the abundant pastures with a
variety of species and / or cultivars. Flexibility of leaf and stem of cocksfoot (Dactylis
glomerata L.), for example, is positively correlated with palatability amongst cultivars in
autumn (Mizuno, 1999). Taking account of those, morphological measurement is also
needed in this investigation.
12
2.4 Increasing clover content of temperate pastures
2.4.1
Clover breeding and sowing
The importance of a higher clover content means that methods are sought to improve the
clover content of pastures. Plant breeders have suggested that new cultivars may increase
the clover content of pastures (Woodfield et al., 2001). However, recent studies of new
versus old cultivars have shown little impact on clover content on dairy production when
compared at the farm scale (Crush et al., 2006). A further approach is to use reduced
sowing rates of competing grasses or using less competitive grasses such as timothy
(Phleum pratense), with the aim to have less competition with the slow establishing white
clover (Cullen, 1958; Hurst et al., 2000). Cullen (1958) demonstrated deleterious effects of
even moderate (10 kg ha-1) sowing rates of ryegrass on the establishment of slow
establishing species such as white clover. Furthermore, sowing method also has a scope for
being reconsidered to promote ‘clover-friendly’ seed beds. Shallow sowing depth (5 mm)
and firm seed beds are recommended for achieving successful emergence of clover (Brock,
2006)
2.4.2
Spatial separation of monocultures
Edwards et al. (2008) and Parsons et al. (2006) summarised a further strategy to increase
clover content, that being the use of spatially separated monocultures of grass and clover in
the same field. The approach is to present plant species in an area ratio that more closely
matches the animal’s perspective of what constitutes an optimal diet. Animal preferences
then ensure that neither species should be grazed to extinction, and that the direct effects of
selective grazing are minimized. Moreover, the selection and sampling costs associated
with grazing a conventionally finely inter-mixed species sward may be minimised
13
(Champion et al., 2004).
Establishing and growing grass and clover separately may also have some advantages in
the ease of fertilising management and weed control, and the improvement of performance
which each individual species owns (Parsons et al., 2006). In a review Edwards et al.
(2008) showed the per-animal performance from spatially separated monocultures of
ryegrass and white clover is generally higher than that from conventional mixtures (c. +
10% in milk; + 25% in live weight gain, Table 2.1), and similar to that in pure white clover
monocultures.
Apart from its great potential, there are some concern to be considered for spatial
separation of pasture species. White clover monocultures produce 25% less in annual
herbage production than a fully fertilised perennial ryegrass monoculture (Harris and
Hoglund, 1977). A simple simulation model for New Zealand dairy farming by Cosgrove
(2005), however, shows milk solids per hectare should improve up to a point where 60% of
pasture allowance was white clover. Thus, the demerit of monoculture use in total annual
DM yield is possibly compensated by improvement of animal production. Relatively even
seasonal distribution of forage can also contribute to compensate the disadvantage of lower
DM production at peak in spring in grass monocultures. According to Chapman and Kenny
(2005), the higher DM production of white clover in summer, when adequate soil moisture
is available, is more profitable than that of growing the same amount of feed in spring in
pasture-based dairy farm systems in southern Australia.
Nitrogen transfer losses from white clover to grass are also concerned both aboveground
(as urine and dung distribution should follow partial preference patterns) and underground
(as clover litter is less likely to be provided to the companion grass species) (Edwards et al.,
14
2008), and it may increase the possibility of N deficiency in grass in spatially separated
monocultures in comparison to finely mixed pastures (Sharp, 2007). Further, potential N
losses from clover monocultures may be greater than ryegrass-white clover mixed swards
(Sharp, 2007), especially in cool seasons (MacDuff et al., 1990). Still, the efficient
utilisation of white clover monocultures has a potential for mitigating the total amount of N
fertiliser use for pastoral farming. Pest attack by nematodes and clover root weevil is a be
concerned (Parsons et al., 2006).
2.4.3
N fixation and uptake
In addition to spatial separation of monocultures, Parsons et al. (2006) suggested other
methods, including altering the trade-off between N fixation and uptake of clover. The
rationale is a notion that soil mineral N derived from the legume increases grass proportion
and inevitably decreases clover proportion in mixed swards. To make clover more
competitive against accompanying grass for acquiring soil mineral N, more aggressive root
systems of clover are required (Parsons et al., 2006).
2.4.4
Alternative grass species
A further possibility is to reduce the preference for the legume species relative to the grass
species (Cosgrove and Edwards, 2007). Altering preference for plant species can have
complex outcomes, especially when the animals revisit the same area and the rate of
replacement of the food resources has been altered, dynamically, by the choices the
animals originally made. Initially showing a strong preference for one species may
increase the proportion of that species in the diet. However, the consumption of that
species may reduce the presence and/or growth rate of it meaning that on subsequent visits
15
there may be less of the preferred item, and more of the less preferred item, available
(Parsons et al., 1991). Thus, it would be necessary to reduce the relative preference for a
preferred item (e.g. legume) to have any consequence of increased nutritive value (e.g.
more legume) in the pasture and diet. Preference tests for grasses and legumes, with
legumes other than white clover, exhibited similar partial preferences to those found with
white clover: 75% (Cosgrove et al., 1997) and 73% (Torres-Rodriguez et al., 1997) for
heifers with Lotus corniculatus; and 73% for sheep with sulla (Hedysarum coronarium L.)
(Rutter et al., 2005b). Thus, there appears to be limited scope for altering preference, and
consequently sward composition, simply by moving to an alternative legume species. The
alternative approach is to use variation in preference of animals for grass species. For
example, the use alternative grass species to perennial ryegrass that have a higher
preference relative to white clover. Thus in mixed pastures, there might be less impact on
white clover through selective grazing. An alternative way of expressing this is that plant
species with a lower preference differential are sought and that partial preference closer to
50:50 (clover:grass) than the 70:30 would be advantageous.
There are some reports that certain grass species are more preferred than the others. For
example, Phillips and Yousseff (2003) showed that timothy and perennial ryegrass (cultivar
or ploidy was not mentioned) were more preferred than cocksfoot and red fescue (Festuca
rubra). However, partial preference for white clover was not recorded in this study.
Anecdotal reports also suggest that tetraploid perennial ryegrass is more preferred than
diploid. But, again, no direct tests relative to white clover of whether partial preference
altered have been reported. Meanwhile, other grass species may have a negative effect on
white clover as they are of low preference. Previous work suggests cocksfoot has low
grazing preference (Edwards et al., 1993), although this may be due to the low N content
of cocksfoot herbage.
16
2.5 Compatibility of grass pasture species with white clover in mixtures
It is important to know the compatibility between a grass species and white clover when a
binary mixture of pastures is introduced onto a pastoral farm. Perennial ryegrass generally
has a notorious reputation as a companion species with the clover in terms of low clover
content (Chapter 1). Fothergil and Davies (1993) demonstrated considerably higher white
clover content in tetraploid than diploid perennial ryegrass/white clover mixtures (13.7%
vs 8.3% for tetraploids and diploids, respectively) under set－stocked condition, and
Davies (1988) reported higher level of compatibility of tetraploids with white clover than
diploids.
Unlike these ryegrass species, timothy is regarded as a slow establishing species (230 ℃d
for 50% field emergence, (Moot et al., 2000)) and a comparable plant to white clover
(150 ℃d for 50% field emergence, (Moot et al., 2000)) within a mixed sward. The clover
composition in timothy-white clover mixtures usually exceeds 20%, and is often double of
that found in
ryegrass-based mixtures (Watkin, 1975; Stevens et al., 1993; Charlton and
Stewart, 2000). Moreover, Skuodiene and Repsiene (2005) achieved over 30% clover
content in timothy-white clover swards, with the white clover composition reaching 38.5%
within the same type of swards in the cutting and rotational grazing system.
Cocksfoot is also known as a slow establishing (Moot et al., 2000) and as an aggressive
species in dryland pasture (Moloney, 1993). This implies that cocksfoot may be compatible
with white clover during the establishment phase in mixed pastures but incompatible after
establishment. Black and Lucas (2000) reported that white clover proportion in
cocksfoot-white clover pastures declined from 24% at 15 months after sowing to 1% at 3
17
years, and the compatibility of cocksfoot with white clover was higher during the
establishment phase than diploid perennial ryegrass but lower after the pasture developed.
In all these studies, however, it is not clear whether differences in composition are due to
these species having a high grazing preference or poor competitive ability during
establishment or as adult plant relative to white clover.
2.6 Aim
The aim of this thesis is to consider the potential role of using alternative grass species to
alter partial preference with the goal of increasing clover content of pastures. Two studies
were conducted. The first compared growth, composition and morphology of pastures
made up of white clover growing with either diploid perennial ryegrass, tetraploid
perennial ryegrass, timothy or cocksfoot. The second examined sheep partial preference for
white clover compared to same four grass species.
2.7 Specific objectives of this study were:
1. To determine the effect of grass species on pasture production, composition and
morphology of mixed grass-white clover pastures over the establishment year; and
2. To determine the effect of grass species on partial preference of sheep for white clover.
18
3 CHAPTER THREE
THE EFFECT OF GRASS SPECIES ON PASTURE PRODUCTION,
COMPOSITION,
SEEDLING
ESTABLISHMENT
AND
MORPHOLOGICAL PROPERTY OF MIXED GRASS-CLOVER
PASTURES
3.1 Materials and methods
3.1.1
Experimental site and design
The study was carried out in plots established in Iversen field I1, Field Services Centre (43
º 38 ’S, 172 º 28 ’E, 11 m a.m.s.l), Lincoln University in January 2007. The soil type is a
Wakanui silt loam (Udic Ustocherept, USDA Soil Taxonomy). The experimental design
was four replicates of grass-clover mixtures with four treatments of grass species: (i)
tetraploid perennial ryegrass, (ii) diploid perennial ryegrass (Lolium perenne L.) (both
ryegrass cultivars contained AR1 endophyte), (iii) timothy (Phleum pratense L.) and (iv)
cocksfoot (Dactylis glomerata L.), sown with white clover (Trifolium repens L.) and laid
out in a randomised block design giving 16 plots (Figure 3.1). The dimension of each
experimental plot was 4.2 m x 5 m.
Results of soil tests taken across the experiment on
20 January 2007 are shown in Table 3.1 (MAF Quick Test). Based on these results,
superphosphate was applied to all plots on 20 March 2007 at 300 kg superphosphate ha-1.
Table 3.1 Soil pH, Olsen P and calcium, magnesium, sodium and sulphur (all MAF quick
the units) for experimental site on 20 January 2007.
pH
6.4
Olsen-soluble P
(
10
㎍ ml )
Ca
Mg
K
Na
13
32
10
14
-1
Sulphate-S
5
19
Rep 1
Rep 2
Tim
CF
DipRG
TetRG
DipRG
Tim
CF
TetRG
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
CF
DipRG
TetRG
Tim
TetRG
DipRG
Tim
CF
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
5m
(Plot No.)
5m
Rep 3
Rep 4
4.2m
Figure 3.1 Plot design for the pasture production, composition, seedling establishment and
morphological study. Codes used: Tim: timothy; DipRG: diploid perennial ryegrass;
TetRG: tetraploid ryegrass; CF: cocksfoot. All plots were sown with white clover.
3.1.2
Pasture establishment
The site was ploughed and rolled, and 2 L ha-1 of glyphosate was applied to remove weeds
on 1 November 2006, and left fallow until January 2007. On 4 January 2007, the site was
again sprayed with 3 L ha-1 of glyphosate, and rolled with a 12 t roller to create a firm seed
bed in preparation for sowing. Seeds were drilled with a 15 cm spacing using a double disc
type drillseeder between each row on 8 January 2007. The species, cultivars and sowing
rates used in the trial are shown in Table 3.2.
Table 3.2 Pasture species, cultivars and sowing rates used in the DM production,
composition and morphology trial.
Pasture species
Cultivar
Ploidy
Perennial ryegrass*
Perennial ryegrass*
Timothy
Cocksfoot
White clover
Quartet
Bronsyn
Viking
Vision
Demand
Tetraploid
Diploid
Diploid
Diploid
Diploid
Sowing rate
Seed No.
(kg ha )
(seeds m )
Thousand seed
weight (g)
15
12
3
3
3
460
543
813
500
446
3.264
2.208
0.369
0.600
0.673
-1
-2
*Ryegrass cultivars contained ARI endophyte.
20
3.1.2.1 Irrigation
The plots were irrigated on 3 occasions during the experiment on 27 February, 29 March
and 26-27 November in 2007. Approximately 28 mm was applied each time.
3.1.2.2 Grazing
Pastures were topped with a mower to approximately 4 cm height on 16 March 2007.
Thereafter, pastures were grazed in common at approximately 1-2 month intervals under
mob-grazing conditions with around 100 ewe hoggets for all four replicates (336 ㎡) or 50
to 70 head for two replicates (168 ㎡). Grazing commenced once pasture mass had
reached around 2500 kg DM ha-1 on the plots with the highest DM (Plate 3.1), and grazing
ceased when pasture mass reached on the lowest plots with around 1000 kg DM ha-1 (Plate
3.3). Grazing typically took 1 to 2 days depending on pasture mass (Plate 3.2). There was
some uneven grazing, reflecting differences in grazing preference. To minimise these
impacts and even up pasture height, pastures were trimmed with a mower to 4 cm high
within 1 to 2 days post grazing.
21
Plate 3.1 Photo showing experimental plots in August 2007.
Plate 3.2 Photo showing experimental plots during grazing in August 2007.
22
Plate 3.3 Photo showing a diploid perennial ryegrass plot after grazing in August 2007.
3.1.3
Measurements
3.1.3.1 DM production and composition
Pasture mass and composition were measured in two 0.2 m2 quadrats cut to ground level
before and after each grazing which occurred on 15 March, 28 March, 25 May, 24 August,
28 September, 7 November and 6 December in 2007. From the quadrat cut before grazing,
a sub-sample was sorted into live grass, white clover, weeds and dead material. The sorted
samples were oven-dried at 65˚C for at least 24 hours, weighed and the percentage
contribution of the different components determined. From post-grazing quadrats, only
pasture mass was measured. Accumulated DM between successive harvests was used to
calculate total DM production and pasture growth rates.
23
3.1.3.2 Grass and white clover populations
The number of grass and white clover plants within a unit length of 40 cm was counted at
four points in each plot positioned along a drill row on 2 March 2007 and 9 January 2008.
To allow comparison with an area with no grass competition, the number of white clover
plants was counted in four 0.4 m drill rows in 3 replicates of the preference trial plots on
the same date (For details of the these plots, see clover plots for preference trial; Figure 4.1
and Section 4.1.1).
3.1.3.3 Grass and clover plant morphology
Morphological measurements of grass and white clover plants were taken on 12 March
2007 (i.e. approximately 2 months after sowing and 4 days before the first harvest) and 9
January 2008 (i.e. approximately 1 year after sowing and 1 month after the final grazing
occasion). At each sampling date, a total of 10 white clover and 5 grass plants were
collected from each plot. Root material was collected with each sample by digging plants
to a depth of approximately 20 to 25 cm. Soil was then washed from each plant. For each
clover plant, the number of stolons and fully expanded leaves were counted, and the height
of the tallest petiole was measured. A stolon was defined as, and counted, when it had an
expanded leaf on a piece of visible stolon. An expanded leaf was recognised as a pair of
those in a condition that these leaves on a single petiole had more than two leaflets
expanded with a 150 degrees or more angle. For each grass plant, the height of the longest
extended tiller was measured, and the number of tillers was counted. Plants were then
separated into root and shoot material, and the bulked samples from each plot oven-dried at
65 ºC for 24 hours before weighing. The mean root and shoot weight per plant was then
24
calculated. Of these morphological measurements, tallest shoot length, root mass and
number of opened leaves of clover were measured to see the influence of companion grass
species on the growth of clover among treatments at the early and late phase of the first
year of establishment.
3.1.3.4 Pasture nutritive value
Pasture snip samples designed to be representative of the material eaten by sheep (top third
of canopy) were collected from each plot prior to grazing in November 2007 (7 hours after
sunrise, one day before the test on each occasion) and then pooled across four replicates.
Samples were freeze-dried and ground, before analysis for the concentration of
metabolisable energy (ME) by near infra-red spectrometry (NIRS).
3.1.4
Calculations
3.1.4.1 Total DM yield of pasture, grass, clover, weeds and dead material from sowing to
the end of grazing trials
Total DM yield was computed to be the sum of the differences in herbage mass between
post grazing mass and the pre grazing mass at the following grazing, and the herbage mass
before topping on 16 March 2007.
3.1.4.2 Daily growth rates of pasture, grass, clover and weeds at each grazing time
Daily growth rate of pasture, grass, clover and weeds were calculated respectively using
the differences of herbage DM mass of these components between each pre grazing at a
25
trial and the post-grazing at the very previous grazing divided by the number of days
between each trial.
3.1.4.3 Botanical composition
The proportion of grass, clover, weeds and dead material were determined by calculating
the % on a dry matter mass basis of these herbage components in the sub-sample from the
quadrat sample collected before each grazing trial.
3.1.5
Weather data
Monthly rainfall and air temperature were collected from the Broadfields weather station,
located approximately 2 km from the experimental site.
3.1.6
Statistical analysis
Data were analysed by analyses of variance (ANOVA) of a randomized block design using
the statistical package GenStat (2007, GenStat 10th Edition, Lawes Agricultural Trust,
Rothamsted experimental Station). Means were separated by LSD test following a
significant ANOVA. For total DM yield, daily growth rates and botanical composition data,
all analyses were carried out on the average of the two samples for each plot before these
statistical tests.
26
3.2 Results
3.2.1
Climate
Monthly rainfall and average monthly air temperature at Lincoln University during the
field experiment and in the past 20 years are shown Tables 3.3 and 3.4. Monthly rainfall
was less than average from January to May 2007 except March (magnitude of difference of
rainfall is 16.8 mm). After May, monthly rainfall was higher than average until the end of
2007, and the average of monthly rain fall over that period was more than 20 mm higher
than that in the past. Total precipitation during this research (661.9 mm) was slightly
higher than the average of last 20 years (620.6 mm) (Table 3.3). Air temperature was
warmer than average in May 2007 and cooler than average in December 2007.
27
Table 3.3 Monthly rainfall at Lincoln University from January 2007 to January 2008. The
20 year average between 1986 and 2006 is shown for comparison.
Monthly rainfall (mm)
Long-term
average
monthly rainfall (mm)
2007
Jan
16.4
41.3
Feb
Mar
Apr
15.4
49.6
36.3
36.3
44.2
49.4
May
Jun
Jul
Aug
21.2
61.6
62.4
100.0
51.8
58.1
52.0
60.2
Sep
Oct
Nov
3.0
97.6
68.6
39.7
48.9
48.8
Dec
2008
Jan
110.6
48.7
19.2
41.3
Total
Average
661.9
50.9
620.6
47.7
28
Table 3.4 Average monthly air temperature at Lincoln University from January 2007 to
January 2008. The 20 year average 1986 and 2006 is shown for comparison.
Average monthly air temperature
(℃)
Long term average monthly air
temperature
(℃)
2007
Jan
15.0
16.4
Feb
Mar
Apr
May
15.7
15.6
11.4
11.8
16.3
14.5
11.7
9.5
Jun
Jul
Aug
5.7
6.4
7.3
6.7
6.0
7.3
Sep
Oct
Nov
10.9
11.3
12.7
9.4
11.5
12.8
Dec
2008
Jan
12.6
15.1
17.1
16.4
Average
11.8
11.8
3.2.2
Total DM yield
Total DM yield of pasture, grass, clover, weeds and dead material from January 2007 to
December 2007 is shown in Table 3.5. Tetraploid and diploid ryegrass plots had higher
total DM than timothy and cocksfoot (Table 3.5). Total DM yield of grasses ranged from
2182 DM ha-1 for timothy to 9562 DM ha-1 produced for tetraploid ryegrass. Total grass
DM production of tetraploid- and diploid ryegrass was approximately three times that of
timothy and cocksfoot. Total white clover DM production was greater by 4 times in
timothy, followed by cocksfoot, compared with tetraploid ryegrass and diploid ryegrass,
with significant differences between each species (P < 0.001). Total weed DM production
was 3-4 times greater in timothy and cocksfoot than diploid and tetraploid ryegrass (P <
29
0.001). In contrast, total dead DM production in timothy and cocksfoot was only 40% of
that in diploid or tetraploid ryegrass (P < 0.001).
Table 3.5 Total dry matter production of pasture, grass, clover, weed and dead material
components (kg DM ha-1) from 8 January to 6 December 2007. P-values are from ANOVA.
Means were separated using a least significant difference (LSD) following a significant
ANOVA. Values followed by different letters within a column are significantly different.
Total
Grass
Clover
Weeds
Dead material
Tet RG
12521a
9562 a
1310c
201 b
1448a
Dip RG
11733a
9158 a
818 d
93b
1663a
Timothy
9751b
2182 b
5936a
771 a
863 b
Cocksfoot
9654b
2679 b
5311b
741 a
923 b
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
999
576
430
213
101
P-value
LSD
3.2.3
Pasture growth rates
Pasture growth rates of the mixed grass-clover pastures from 8 January 2007 to 6
December 2007 are shown in Figure 3.2. Pasture growth rate ranged from 6 kg DM ha-1 d-1
in diploid ryegrass in winter to 91 kg DM ha-1 d-1 in tetraploid ryegrass in the summer
(December) 2007. Pasture growth rate at the first three harvests was greater in tetraploid
and diploid ryegrass than timothy and cocksfoot. Pasture growth rate declined in all
pastures during winter, and did not differ significantly among grasses in September.
Between September and October, timothy and cocksfoot pastures grew faster (P<0.05) than
tetraploid and diploid ryegrass pastures. Between October and November, diploid ryegrass
had slower growth (nearly 14 kg DM ha-1 d-1) than the other pastures (P<0.001).
30
140
*
*
***
*
***
-1
Daily growth rate (kg DM ha d )
120
-1
100
Tetraploid ryegrass Trt.
Diploid ryegrass Trt.
Timothy Trt.
Cocksfoot Trt.
80
60
40
20
0
Mar
May
Jul
Sep
Nov
Jan
Grazing date
Figure 3.2 Pasture growth rates of grass-white clover pastures from 8 January 2007 to 6
December 2007. The vertical bars represent the LSD for comparing the differences among
grass treatments at each date. Asterisks indicate significance of treatments effects at each
date, * P < 0.05, *** P < 0.001(***).
3.2.4
Botanical composition
Tetraploid and diploid ryegrass pastures had around four times higher (P<0.001) grass
percentage almost throughout the trial than timothy and cocksfoot (Figure 3.3a). There was
no significant difference between diploid and tetraploid ryegrass in grass percentage,
except in November, when diploid ryegrass had a higher grass percentage. There was no
significant difference between timothy and cocksfoot in grass percentage throughout the
year. The percentage of clover showed the reverse pattern being greater (P<0.001) in
timothy and cocksfoot than tetraploid and diploid ryegrass throughout the year (Figure
3.3b). No significant differences in clover percentage were found between tetraploid and
31
diploid ryegrass or between timothy and cocksfoot at any date. The dicot weed content of
the pastures (mainly Capsella bursa-pastoris, Poa spp. and Rumex spp.) fluctuated,
ranging from nearly 0% in June to 30% in September (Figure 3.3c). Weed percentage was
not significantly different among the four pastures, except in May where there was a
significantly (P<0.05) higher percentage (7.8%) of weeds in timothy than tetraploid and
diploid ryegrass and cocksfoot. The percentage of dead material ranged from 0 to 15.5%
(Figure 3.3d). In March and April, dead material percentage was higher (P<0.001) in
tetraploid and diploid ryegrass than timothy or cocksfoot pastures.
32
(a) Grass
140
***
120
***
***
***
***
***
***
***
***
Grass (%)
100
80
60
40
20
0
(b) Clover
140
120
***
*** ***
***
***
Clover (%)
100
80
60
40
20
0
(c) Weed
140
120
Weeds (%)
100
80
60
***
*
40
20
0
(d) Dead material
140
Dead material (%)
120
Tetraploid perennial ryegrass Trt.
Diploid perennial ryegrass Trt.
Timothy Trt.
Cocksfoot Trt.
100
80
60
40
***
*
***
*
***
*
20
0
Mar
May
Jul
Sep
Nov
Jan
Grazing date
Figure 3.3 The botanical composition (% by DM in pre-grazing mass) of (a) grass, (b)
white clover, (c) weeds (dicot and grass) and (d) dead material in the four grass-white
clover mixtures from March to December 2007. The vertical bars represent the LSD for
comparing the differences among swards types at a given grazing occasion. Asterisks
indicate significance of treatments effects at each date, * P < 0.05, *** P < 0.001(***).
33
3.2.5
Plant density
Table 3.6 shows the number of grass and clover plants per ㎡ in March 2007 and January
2008 in each pasture, and the number of clover plants in pure swards of white clover (see
preference experiment, Chapter 4, Section 4.1.1 and 4.1.2). On average, the combined total
number of grass and clover plants per ㎡ declined during the experiment from an average
of 158 to 118. Total plant number did not differ among treatments in the early phase of
pasture development (March 2007) but was greater (P<0.05) in tetraploid and diploid
ryegrass than timothy and cocksfoot in January 2008. In March 2007 and January 2008
there were more grass plants in diploid and tetraploid ryegrass than timothy and cocksfoot.
The decline in plant number from March 2007 to January 2008 of grasses ranged from
25% (for both ryegrasses) to 35% for timothy. In March 2007, the number of clover plants
was highest in cocksfoot, followed by timothy, with the fewest clover plants in tetraploid
and diploid ryegrass pastures. In January 2008, the number of clover plants was highest
(P<0.05) in timothy, followed by cocksfoot, tetraploid ryegrass with diploid ryegrass the
lowest (Table 3.6). The number of clover plants declined the least in tetraploid ryegrass (58
versus 55 plants m-2) and the most in cocksfoot (96 versus 63 plants m-2). Grass-white
clover plots had 42 to 70% of the number of clover plants in pure swards in January 2008.
34
Table 3.6 Number of plants per ㎡ in mixed swards and pure monoculture swards in
autumn (March) 2007 and in summer (January) 2008. P values are from one way ANOVA;
ns = no significant difference. Means followed by the same letters are not significantly
different according to LSD tests (α=0.05) following a significant ANOVA.
March2007
Total plants
Grass
Clover
Tet RG
154.7
95.9a
58.8c
Dip RG
160.9
107.2a
53.6c
Timothy
155.7
74.2b
81.5b
Cocksfoot
151.6
54.7b
96.9a
ns
< 0.05
< 0.001
25.4
25.4
14.88
Clover
in monoculture
126.5
P-value
LSD
January 2008
Total plants
Grass
Clover
Tet RG
127.9a
72.2a
55.7bc
Dip RG
122.7a
80.4a
42.3c
Timothy
119.6b
48.5b
71.2a
Cocksfoot
101.1b
37.1b
63.9ab
P-value
< 0.05
< 0.001
< 0.05
LSD
14.4
12.9
13.8
3.2.6
Clover and grass morphology
Clover
in monoculture
100.4
Diploid and tetraploid ryegrass plants had more tillers per plant, longer shoots, and greater
root and shoot mass than timothy and cocksfoot plants in March 2007 (P < 0.05, Table 3.7).
In January 2008, tiller numbers were similar in tetraploid, diploid ryegrass and cocksfoot.
Tetraploid and diploid ryegrass had greater tiller numbers than timothy. Grass shoot mass
in January 2008 was greater in tetraploid and diploid ryegrass and cocksfoot than timothy.
35
Table 3.7 The number of tillers per plant, tallest shoot length, and root and shoot mass of
individual grass species in March 2007 and in January 2008. P values are from one way
ANOVA; ns = no significant difference. Means followed by the same letters are not
significantly different according to LSD tests (α=0.05) following a significant ANOVA.
M arch 2007
No. of tillers
Tallest shoot
length
(mm)
Root mass
(g)
Shoot mass
(g)
Root mass
/shoot mass
Tet RG
10.2a
263a
0.428a
1.12 a
0.38
Dip RG
15.9a
321a
0.411a
1.40 a
0.29
Timothy
4.02 b
103b
0.047b
0.15b
0.31
Cocksfoot
3.65 b
92 b
0.045b
0.11b
0.41
P-value
< 0.05
< 0.05
< 0.05
< 0.05
LSD
5.29
153
0.291
0.78
January 2008
No. of tillers
Tallest shoot
length
(mm)
Root mass
(g)
Shoot mass
(g)
Root mass
/shoot mass
Tet RG
56.9a
220b
3.78
18.5 a
0.20
Dip RG
67.6a
239b
2.43
15.9 a
0.15
Timothy
26.1 b
243b
0.46
6.6b
0.07
Cocksfoot
46.5ab
366a
1.28
16.0 a
0.08
P-value
< 0.05
< 0.001
ns
< 0.05
LSD
26.3
44.0
2.73
8.2
36
In March 2007, the number of stolons per white clover plant and fully opened leaves per
white clover plant were greater in timothy and cocksfoot than tetraploid and diploid
ryegrass pastures. Root mass followed a similar pattern, although cocksfoot and diploid
ryegrass did not differ significantly. In January 2008, stolon and leaf number did not differ
among treatments. Root and shoot mass of clover were lower in tetraploid and diploid
ryegrass pastures than timothy and cocksfoot pastures (Table 3.8).
Table 3.8 The number of stolons, fully opened leaves, length of longest above-ground part,
and root and shoot weight per plant of white clover plants in March 2007 and January 2008.
P values are from one way ANOVA; ns = no significant difference. Means followed by the
same letters are not significantly different according to LSD tests (α=0.05) following a
significant ANOVA.
M arch 2007
No. of stolons
No. of
opened leaves
Tallest shoot length
(mm)
Root mass
(g DM)
Tet RG
2.55 b
9.62b
128
0.0362 c
0.219
0.17
Dip RG
2.80 b
10.52b
140
0.0446 bc
0.282
0.16
Timothy
5.37 a
16.77a
77
0.0751 a
0.391
0.19
Cocksfoot
4.85 a
14.27ab
61
0.0675 ab
0.269
0.25
< 0.05
< 0.05
ns
< 0.05
ns
1.48
4.71
80
0.0253
0.178
P-value
LSD
Shoot mass Root mass
(g DM)
/shoot mass
January 2008
No. of stolons
No. of
opened leaves
Tallest shoot length
(mm)
Root mass
(g)
Tet RG
29.9
28.4
105.3b
0.194b
1.47b
0.13
Dip RG
21.6
24.1
117.4b
0.173b
1.20b
0.14
Timothy
33.3
27.5
158.1a
0.415 a
2.70a
0.15
Cocksfoot
34.6
26.0
131.3ab
0.392 a
2.23ab
0.18
ns
ns
< 0.05
< 0.05
< 0.05
9.8
10.4
35.6
0.125
0.96
P-value
LSD
Shoot mass Root mass
(g)
/shoot mass
37
3.2.7
Pasture nutritive value
Metabolisable energy (ME) of grass and white clover in mixed swards did not differ
among the treatments in November 2007 but values were considerably high (Table 3.9)
Table 3.9 Concentrations of metabolisable energy (MJ ME kg-1 DM) of grass and white
clover in the binary mixtures on 7 November 2007.
Grass
Clover
Tet RG
12.8
13.0
Dip RG
12.3
13.0
Timothy
12.6
13.1
Cocksfoot
12.5
13.1
3.3 Discussion
The total pasture yields and composition data measured over the first year (Table 3.5)
indicate that highly productive pastures of high nutritive value were established in all four
grass treatments. However, the yield and botanical composition were all influenced by the
grass species planted.
3.3.1
White clover yield: timothy and cocksfoot versus ryegrass species
The white clover content was higher in timothy and cocksfoot pastures than tetraploid and
diploid perennial ryegrass swards throughout the trial period. Over the entire growth period
of 11 months 5-6t DM ha-1 of white clover was grown in timothy and cocksfoot swards but
only 0.8-1.3t DM ha-1 in ryegrass swards. The greater white clover content in timothy than
38
diploid ryegrass pastures is consistent with previous research. Watkin (1975) showed white
clover content of timothy-white clover mixtures was about one and a half times that of
diploid ryegrass-white clover pastures.
The higher white clover content of cocksfoot than ryegrass pastures is in contrast to several
other studies noting less clover when white clover is grown with cocksfoot (Moloney,
1993; Black and Lucas, 2000; Hurst et al., 2000), particularly in dryland, sites subject to
summer moisture stress. The difference may reflect that this trial was irrigated and of short
duration. For example, Black and Lucas (2000) in an unirrigated trial on the same soil type
in a neighbouring paddock to this experiment noted similar white clover content in
ryegrass and cocksfoot pastures in the first year, but more white clover in ryegrass than
cocksfoot (albeit low in both) two years later.
The combination of species-specific thermal times (Tt) for seedling emergence and
seedling growth may provide an explanation for the higher clover content in timothy and
cocksfoot than ryegrass pastures, particularly at the initial harvests. For example, white
clover (150 ℃d) and perennial ryegrass (160 ℃d) have similar thermal time requirements
for field emergence (Moot et al., 2000) and would therefore have emerged at similar time.
Thermal time requirements of timothy (230 ℃d) and cocksfoot (250 ℃d) are higher than
that of white clover (150 ℃d) (Moot et al., 2000), meaning white clover would have
emerged earlier. This is reflected in fewer seedlings of timothy and cocksfoot at the first
measurement period.
Differences in the numbers of seeds sown per m-2 were unlikely to
have played a significant role in the number of seedlings of cocksfoot and timothy; the
number of seeds sown was higher in timothy than other species (Table 3.2), while similar
numbers of cocksfoot and ryegrass seeds were sown (Table 3.2).
39
A further explanation for the higher clover content in timothy and cocksfoot pastures may
be the higher grazing preference for these legume dominated grazing plots leading to lower
grazing residuals. Grazing residuals were lower in timothy and cocksfoot (276 and 197 kg
DM ha-1, respectively) than both ryegrasses (531 and 843 kg DM ha-1, respectively for
tetraploids and diploids) (Table 3.10). This may have suppressed timothy and cocksfoot, or
alternatively, provided more space for stolon expansion in timothy and cocksfoot pasture
leading to greater clover content.
Table 3.10 Total dry matter of overall pasture after each grazing occasion (kg DM ha-1)
from 8 January to 6 December 2007.
APR
MAY
SEP
OCT
NOV
DEC
mean
Tet RG
634
1001
410
318
497
327
531
Dip RG
1065
1262
658
674
761
639
843
Timothy
150
318
152
278
176
110
197
Cocksfoot
134
419
264
410
234
195
276
Despite similar thermal time requirements for emergence of white clover (150 ℃d) and
perennial ryegrass (160 ℃d) (Moot et al., 2000), diploid and tetraploid ryegrass were
dominant in clover-mixtures throughout the year (Table 3.5). One of the reasons for this is
that white clover is sensitive to the presence or density of its neighbouring species such as
perennial ryegrass (Haggar et al., 1985). This may be more significant in the early phase of
establishment, rather than the late phase. Some morphological data are consistent with this
possibility in this experiment. For instance, number of stolons and open leaves, and root
mass of white clover were higher in timothy and cocksfoot treatments than diploid and
tetraploid perennial ryegrass treatments at the early stage of establishment (Table 3.8)
whereas tiller number and root mass of grass were higher in both types of perennial
ryegrass-based mixtures than in timothy- and cocksfoot-based mixtures at the same time
40
(Table3.7). Moreover, timing of axillary shoot development is also regarded as a major
determinant of successful ryegrass establishment in clover-mixed swards (Black et al.,
2002). For example, the thermal time requirements for the appearance of the first tiller of
perennial ryegrass is 373 ℃d. This indicates the initiation of secondary leaf development
is considerably earlier than the first white clover stolon (532 ℃d). As a consequence,
ryegrass can expand tillers much earlier than the clover extends their stolons (Black et al.,
2006), meaning that the perennial ryegrass is very competitive for light with white clover.
This was reflected in lower white clover plant weights at the final harvest in both
ryegrasses compared with cocksfoot and timothy, despite there being no differences at the
first harvest (Table 3.8).
3.3.2
White clover content of diploid versus tetraploid perennial ryegrass
There were few differences in clover content between diploid and tetraploid ryegrass
(Figure 3.3b). This contrasts with several previous studies noting higher clover content in
tetraploid than diploid ryegrasses. For example, Fothergill and Davies (1993) demonstrated
higher annual clover content in tetraploids (13.7%) than diploids (8.3%) throughout a
5-year measurement under set-stocked grazing. Swift et al. (1992) also showed high clover
contribution to tetraploid ryegrass mixtures, ranging from 13% to 26% during 3
consecutive years of measurement in a continuous grazing system although diploid mixed
swards were not examined simultaneously. The lack of difference may reflect two main
factors: (i) the short-term nature of mixed swards and (ii) rotational grazing management.
Most of these previous studies were longer than the 12 months of this study, and the
difference in clover composition between grass treatments emerged over time, which
perhaps depends on the difference of tiller numbers and competitive ability between these
grasses (Frame and Boyd, 1986; Fothergil and Davies, 1993; Swift et al., 1993). Grazing
41
management is also likely to alter intake behaviour of ruminants such as selective grazing
(Cosgrove and Edwards, 2007) and consequently affect botanical composition in mixed
swards. The result that mixtures with tetraploid ryegrass were grazed more compared to the
mixtures with diploid ryegrass (post-grazing mass : diploids = 843 kg DM ha-1, tetraploids
= 531 kg DM ha-1, Section 3.3.1) may be a possible explanation to support the idea,
leading to the expectation (Chapter 4) of lower partial preference for clover in the
tetraploid, and less selective pressure. However, in this study, hard rotational grazing may
have minimised the animals’ opportunities for selective grazing and encouraged clover in
both tetraploids and diploids. Greater difference might be expected under a set-stocking
system where the effect of grazing preference and diet selection can be exerted more
strongly.
3.3.3
Total Pasture Production
The greater total production in the ryegrass pastures than cocksfoot and timothy pastures
(2425 kg DM ha-1) over the establishment year was due to the rapid ryegrass establishment
and subsequent growth over the first 3 months. Of particular note, was the greater growth
of the ryegrass pastures than timothy and cocksfoot over the first autumn period. However,
spring growth rates were similar (or occasions higher in cocksfoot) across grass treatments,
indicating limited long term effect.
42
(a)
(b)
Plate 3.4 Seedlings in a tetraploid perennial ryegrass / white clover sward 45 days after
sowing (a and b). (22 February 2007, 10 days before 1st plant number measurement).
43
(a)
(b)
Plate 3.5 Seedlings in a diploid perennial ryegrass / white clover sward 45 days after
sowing (a and b). (22 February 2007, 10 days before 1st plant number measurement).
Diploid ryegrasses grew relatively faster to small extent at this stage in comparison to
tetraploid ones. Clover grew vertically according to the height of neighbouring grass
plants.
44
(a)
(b)
Plate 3.6 Seedlings in a timothy / white clover sward 45 days after sowing (a and b). (22
February 2007, 10 days before 1st plant number measurement). A higher predominance of
clover over timothy. Clover plant height is quite low and has much lesser competition with
timothy.
45
(a)
(b)
Plate 3.7 Seedlings in a cocksfoot / white clover sward 45 days after sowing (a and b) (22
February 2007, 10 days before 1st plant number measurement). An extremely low number
of cocksfoot plants.
46
(a)
(b)
Plate 3.8 Tetraploid ryegrass / white clover mixture (above, a) and timothy / white clover
mixture (below, b) on 20 Jun 2007, six months after sowing.
47
4 CHAPTER FOUR
THE EFFECT OF GRASS SPECIES ON GRAZING PREFERENCE OF
SHEEP FOR WHITE CLOVER AND GRASS
4.1 Materials and methods
4.1.1
Experimental site and design
The grazing study was carried in plots established in field I1, Field Services Centre, (43 º
38 ’S, 172 º 28 ’E, 11 m a.m.s.l), Lincoln University in January 2007. The soil type is a
Wakanui silt loam (Udic Ustocherept, USDA Soil Taxonomy). The experimental design
was three replicates of four grass species laid out in a randomised block design. The
experimental plots were 4.2 m x 10 m swards of pure grass sown beside 4.2 m x 10 m
swards of pure white clover (Figure 4.1). Soil samples were collected on 20 January 2007
and analysed with a MAF Quick Test (see Table 3.1). Based on these results,
superphosphate (0-9-0-12) was applied to the plots on 20 March 2007 at 300 kg ha-1.The
grass monocultures excluding cocksfoot and clover plots were fertilized with 50 kg N ha-1
as urea on 20 March 2007.All plots were fertilised with 40 kg N ha-1 as urea on 6 August
2007. White clover plots were fertilized as well as pasture growth at this time was slow due
to cool temperatures. On 27 September and 26 October, 20 kg N ha-1 as urea was applied
only onto the grass monocultures. The aim of fertilizing the grass monocultures throughout
with N fertiliser was to establish an N concentration in the herbage that would approximate
that of green, leafy grass growing within a grass-clover mixture.
48
Rep 1
Tim
(1)
WC RGDip WC
(2)
(3)
(4)
Rep 2
CF
(5)
WC RGTet WC
(6)
(7)
(8)
CF
WC
Tim
WC RGDip WC RGTet WC
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
Rep 3
RGDip WC
CF
WC RGTet WC
Tim
WC
(19)
(20)
(23)
(24)
10 m
(17)
(18)
(21)
(22)
4.2m
Figure 4.1 Plot design for grazing preference trial. Tim: timothy; DipRG: diploid perennial
ryegrass; TetRG: tetraploid ryegrass; CF: cocksfoot; WC: white clover.
4.1.2
Pasture management
4.1.2.1 Cultivars and sowing rates
The site was ploughed and rolled, and 2 L ha-1 of glyphosate was applied to remove weeds
on 1 November 2006, and left fallow until January 2007. On 4 January 2007, the site was
again sprayed with 3 L ha-1 of glyphosate, and rolled with a 12 t roller to create a firm seed
bed in preparation for sowing. Seeds were drilled with a 15 cm spacing using a double disc
type drillseeder between each row on 8 January 2007. The species, cultivars and sowing
rates used are shown in Table 4.1. To establish sufficiently dense monocultures for the
preference tests, higher sowing rates were set up for monoculture plots compared to
mixture plots. All plots excluding cocksfoot monocultures were topped to even pasture
height on 16 March 2007. Cocksfoot establishment was poor at that time, and then plots
were re-sown with 3 kg cocksfoot ha-1 on 23 March 2007.
49
Table 4.1 Pasture species, cultivars and sowing rates used in the grazing preference trial.
Sowing rate
Seed No.
(kg ha )
(seeds m )
Thousand seed
weight (g)
Tetraploid
20
613
3.264
Bronsyn
D iploid
15
679
2.208
Timothy
Viking
D iploid
6
1626
0.369
Cocksfoot
Vision
D iploid
3
500
0.600
Demand
D iploid
4
594
0.673
Pasture species
Cultivar
Ploidy
Perennial ryegrass*
Quartet
Perennial ryegrass*
White clover
-1
-2
*Ryegrass cultivars contained ARI endophyte.
4.1.2.2 Irrigation
The plots were irrigated on 3 occasions during the experiment on 27 February, 29 March
and 26-27 November in 2007. Approximately 28 mm was applied each time.
4.1.3
Preference tests
Preference tests for the four grass-white clover combinations were carried out on 5
occasions in 2007: 22-24 May, 11-13 September, 15-18 October, 8-12 November and
11-14 December. The rationale of using a range of dates was to cover variation in grass and
clover quality and morphology throughout the year, including those key changes that occur
during the grass reproductive phase.
For preference tests, each plot containing a sward of grass beside a sward of white clover
was stocked with three Coopworth ewe hoggets (mean liveweight : 42.8kg) (i.e. 3 sheep x
4 plots = 12 sheep per block) for 8 hours (from 9 am to 5 pm) only on a test day (Plate 4.1).
The preference test was carried out on one block of four plots per day, thus taking 3 days to
50
complete all three blocks. A different set of 12 sheep (36 in total) was used for each block.
When not in use, the sheep were grazed in a permanent pasture paddock which is mainly
grass-dominant, located on the next paddock prior to the tests from one day before the first
test day to the end of the third test day. Three sheep was considered the minimum number
of sheep to be used in a preference test so that sheep exhibit ordinary grazing behaviour
(Penning et al., 1993). After each preference test, plots were grazed with a large mob of
sheep to ensure a low pasture residual across the pastures of approximately 1000 kg DM
ha-1 and were trimmed with a mower to approximately 4 cm high within 1 or 2 days post
grazing to even up pasture height. Plots were then allowed to regrow without grazing until
the next preference test took place.
Plate 4.1 View of a preference test site from the observation tower (September 2007).
51
4.1.4
Measurements
4.1.4.1 Animal behaviour measurements
Foraging behaviour was scored by visual scan sampling of each sheep at 5 minute intervals
throughout the day 9 am to 5 pm (Orr et al., 2005). Sheep were scored in the following
categories: grazing white clover, grazing grass, ruminating and idling (Plates 4.2 and 4.3).
Idling was defined as when a sheep had no specific jaw movements and was often
associated with the sheep sitting down (Orr et al., 2005) (Plate 4.3). Grazing time on grass
and clover, and ruminating and idling time were converted from the observation scores
multiplied by 5, assuming the same behaviour over the previous 5 minutes. The percentage
of grazing time on clover was calculated as a measure of grazing preference. A selection
coefficient for white clover (θ) was adapted from the method of Ridout and Robson
(1991) and calculated as: θ = (Proportion clover in total dry matter intake / proportion
grass in total dry matter intake) / (Proportion clover in pre-grazing mass / proportion grass
in pre-grazing mass). For this coefficient, a value of 1 for θ implies that no differential
selection is taking place, while a value of 2 indicates that a unit of clover is twice as likely
to be grazed as a unit of grass. Values of below 1 represent that a unit of grass is more
likely to be preferred than a unit of clover. The average of the three sheep was used as the
unit of replication as their behaviour could not be considered independent of each other
(Penning et al., 1993). The percentage of dry matter intake (DMI) for clover was calculated
as: DMI% for clover = (difference between pre-grazing and post-grazing mass in clover) /
[(difference between pre-grazing and post-grazing mass in clover) + (difference between
pre-grazing and post-grazing mass in grass)].
52
Plate 4.2 View of a preference test site from the observation tower (September 2007). Two
sheep grazing on clover and a sheep ingesting grass.
Plate 4.3 A sheep in idle state. This sheep would be scored as “idling” due to no jaw
movement. It will be counted as “ruminating” in the case where jaw movement (chewing
or ruminating) is observed.
53
4.1.4.2 Pasture measurements
Sward surface height (SSH)
Sward surface height was measured in each grass and white clover plot using a sward stick
before and after the preference test and every 2 hours during the grazing preference test,
with 25 contacts per plot. Visual scan sampling ceased while SSH measurements were
being made but sheep remained in the plot. The decline in height and clover during the
preference test was used as a further measure of grazing preference.
Pasture mass and plant morphology
Pasture measurements were made in two 0.2 ㎡ quadrats cut to ground level in each grass
and white clover plot of each replicate prior to and after each preference test. For grasses, a
subsample (around one quarter of sample) was taken and measurements made of: number
of tillers, extended tiller length and sheath tube length of each tiller.
The sub-sample was
then sorted into live lamina, dead material, sheath material and reproductive parts, For
white clover, weeds were separated from pure clover and a subsample (around one quarter
of sample) was taken, and then dried after the number of grazed and ungrazed (lamina
intact) petioles were counted. The dry weight (oven-drying at 65˚C for 24 hours) of the
components and the number of tillers was calculated from the contribution of their weight
to total weight of the respective samples and the total sample weight. The proportion of
grazed petioles was used as an additional measure of grazing preference.
54
Pasture nutritive value
Pasture snip samples designed to be representative of the material eaten by sheep (top third
of canopy) were collected from each plot prior to grazing (7 hours after sunrise, one day
before the test on each occasion). Samples were freeze-dried and ground, before analyses
for the concentration of metabolisable energy (ME), the percentage of N, water soluble
carbohydrate (WSC), digestibility, and neutral detergent fibre (NDF) by near infra-red
spectrometry (NIRS). Snip samples from the May, October and December tests were
analysed for macronutrients (calcium (Ca), potassium (K), magnesium (Mg), nitrogen (N),
sodium (Na), phosphorus (P) and sulphur (S)) by the methods of Dumas combustion for N
and nitric acid / hydrogen peroxide digestion followed by ICP-OES for other elements at R
J Hill Laboratories Limited (Hamilton, New Zealand). Macronutrient samples were pooled
across the three replicates.
4.2 RESULTS
4.2.1
May preference test
4.2.1.1 Pasture characteristics
Grass pasture mass ranged from 1245 to 2637 kg DM ha-1 and was significantly greater in
timothy and cocksfoot than both ryegrasses (Table 4.2). Both ryegrasses had a lower
proportion of grass and higher proportion of dead matter than timothy and cocksfoot (Table
4.2). Cocksfoot plots contained the lowest number of tillers per ㎡ and had the longest
leaf and sheath tube. (Table 4.2). The proportion of lamina was higher in timothy and
55
cocksfoot than both ryegrasses (Table 4.2). Clover had lower pasture mass than all grass
plots, and only contained a small percentage of weed. Cocksfoot was considerably taller
than the other grass species at the start of grazing since only cocksfoot plots were not
topped on 16 March 2007 (see Section 4.1.2.1).
Table 4.2 Pasture mass, composition and morphology of grass and clover plots in each
grass treatment combination prior to the May preference test. P values are from one way
ANOVA. Means with same subscript within a row are not significantly different according
to LSD test following a significant ANOVA.
TetRG
DipRG
Tim
CF
P - value
LSD
1338b
61.2b
1245b
57.4b
2637a
78.7a
2396a
82.8a
0
38.8a
3.6
42.2a
1.0
21.2b
2.3
14.9b
< 0.05
< 0.001
ns
< 0.001
947.1
7.44
3.37
7.69
No. of tillers per m
No. of greenleaves per tiller
Leaf length (mm)
Sheath tube length (mm)
2728a
2.0b
129.4b
33.2b
3191a
1.8b
123.8b
36.1b
3403a
3.4a
130.4b
38.5b
1904b
3.3a
234.3a
80.2a
< 0.01
< 0.001
< 0.001
< 0.001
715.1
0.33
22.13
6.15
Lamina (%)
Psuedostems (%)
Dead matter (%)
Reproductive parts (%)
38.7b
22.5b
38.8a
0
34.8b
22.8b
42.4a
0
53.3a
25.5b
21.2b
0
54.6a
30.1a
15.3b
0
< 0.001
< 0.05
< 0.001
6.39
4.46
7.48
Sward surface height (cm)
15.5
16.7
18.1
28.6
< 0.01
4.36
Clover plots
-1
Total pasture DM yield (kg ha )
Clover (%)
Weed (%)
1252
96.0
4.0
1111
91.7
8.3
1244
92.6
7.4
1287
87.7
12.3
ns
ns
ns
203.2
13.20
13.2
Sward surface height (cm)
9.8
9.1
9.1
9.1
ns
0.74
Grass plots
-1
Total pasture DM yield (kg ha )
Grass (%)
Weed (%)
Dead (%)
2
56
ME (range 12.6 to 13.5 MJ ME kg-1 DM) and N% (range 4.8 to 5.3%) were high in all
forages (Table 4.3). Cocksfoot was considerably lower in WSC content than other species
(Table 4.3). Clover contained markedly lower NDF than grasses (Table 4.3).
Table 4.3 Nutritive value of grass and clover determined by NIRS prior to the May
preference test. P values are from one way ANOVA. Means with same subscript within a
row are not significantly different according to LSD test following a significant ANOVA.
TetRG
DipRG
Tim
CF
Clover
P - value
LSD
ME (MJ ME kg DM)
13.4b
13.0c
13.5a
12.7d
12.6d
< 0.001
11.27
N (%DM)
4.8c
4.8c
5.1ab
5.3a
4.9b
< 0.05
0.27
Water soluble carbohydrate
(%DM)
17.3ab
16.1bc
17.7a
11.0d
15.3c
< 0.001
1.16
Digestibility (%DM)
87.2a
85.0b
87.7a
82.7d
83.4c
< 0.001
0.81
Neutral detergent fibre (%DM)
29.6bc
30.7b
28.6c
34.8a
18.5d
< 0.001
1.90
-1
57
Both ryegrasses had a higher S concentration than cocksfoot and timothy (Table 4.4).
Timothy contained the highest concentration of P, and the lowest concentration of K, Ca
and Na (Table 4.4). Of note is the very low Na concentration in timothy. Clover was higher
in Ca and lower in K and S than grasses (Table 4.4). It also had lower Na content compared
to all grasses except timothy.
Table 4.4 The concentration of minerals (%DM) contained in laminae of grass and clover
prior to the May grazing trial.
Tet RG
Dip RG
Tim
CF
WC
P%
0.35
0.35
0.46
0.32
0.41
K%
4.1
3.9
3.7
4.0
3.4
S%
0.54
0.50
0.34
0.40
0.29
Ca%
0.45
0.48
0.32
0.32
0.99
Mg%
0.20
0.21
0.17
0.17
0.18
Na%
0.37
0.35
0.02
0.11
0.11
4.2.1.2 Grazing behaviour and pasture changes
Total grazing time was longer in diploid perennial ryegrass than other treatments (Table
4.5). Other grazing behaviour and preference measures did not differ among grass species.
Sheep spent 43% to 50% of grazing time on clover. Selection coefficient was higher in
timothy than others but did not significantly differ across the treatments. There were no
significant effects of grass species on decline in the percentage of grazed petioles or the
decline in pasture height or mass (Figure 4.2a and b; Table 4.5). On average grass height
declined 8.2 cm, and clover height declined 3.1 cm.
58
Table 4.5 Grazing behaviour and changes in pasture for the May preference test. P values
are from one way ANOVA. Means with same subscript within a row are not significantly
different according to LSD test following a significant ANOVA. DMI = dry matter intake.
SSH = sward surface height.
TetRG
DipRG
Tim
CF
P - value
LSD
Grazing behaviour
Total grazing (min)
Grazing grass (min)
Grazing clover (min)
Ruminating (min)
Idling (min)
Grazing time on clover (%)
Selection coefficient for clover
DMI for clover in diet (%)
227ab
131
97
76
130
43.0
0.59
63.2
243a
131
112
56
134
46.5
0.50
44.3
198b
103
94
69
167
48.2
1.42
68.1
198b
100
98
84
151
50.2
0.90
48.0
< 0.05
ns
ns
ns
ns
ns
ns
ns
32.5
58.2
46.0
52.3
66.6
23.68
1.352
82.5
Pasture changes
Decline of clover SSH (cm, overall)
Decline of grass SSH (cm, overall)
Decline of clover SSH (cm, 1st 2hrs)
Decline of grass SSH (cm, 1st 2hrs)
Grazed petioles (%)
2.8
7.9
1.2
1.9
55.1
3.6
8.7
0.9
2.2
49.4
2.6
7.6
0.8
1.7
49.8
3.5
8.5
1.0
2.9
61.4
ns
ns
ns
ns
ns
0.83
4.24
1.42
1.75
23.58
45.3
29.8
38.9
42.0
ns
46.42
Decline of clover mass (%)
59
40
(a)
**
**
30
20
Height (cm)
Height (cm)
10
0
14
(b)
12
10
8
6
Height (cm)
4
Tetraploid perennial ryegrass Trt.
Diploid perennial ryegrass Trt.
Timothy Trt.
Cocksfoot Trt.
2
0
9:00
11:00
13:00
15:00
17:00
Time
Figure 4.2 Sward surface height (SSH) (cm) of (a) grass and (b) clover plots in paired
comparisons during the May trial. The vertical bars represent the LSD for comparing the
differences among grass treatments at the start and end of grazing occasion.
60
4.2.2
September preference test
4.2.2.1 Pasture characteristics
Grass pasture mass ranged from 1708 to 2033 kg DM ha-1 but did not differ among
treatments (Table 4.6). Grass, weed and dead percentage did not differ among grass species
(Table). Both ryegrasses had more tillers per ㎡ than timothy and cocksfoot (Table 4.6),
but fewer green leaves per tiller (Table 4.6). Clover had a lower pasture mass than all grass
plots, and only contained a small percentage of weeds.
Table 4.6 Pasture mass and pasture composition of grass and clover plots in each grass
treatment combination prior to the September grazing. P values are from one way ANOVA.
Means with same subscript within a row are not significantly different according to LSD
test following a significant ANOVA.
TetRG
DipRG
Tim
CF
P - value
LSD
Total pasture DM yield (kg ha-1)
Grass (%)
Weed (%)
Dead (%)
1708
89.8
0.6
9.6
1714
88.1
1.4
10.6
2033
92.0
0.8
7.2
1592
87.1
6.7
6.2
ns
ns
ns
ns
727.9
9.15
7.32
6.6
No. of tillers per m2
No. of greenleaves per tiller
Leaf length (mm)
Sheath tube length (mm)
5003a
2.6c
114.9
15.4b
4789a
2.6c
110.7
21.8a
2998b
3.6a
126.7
14.1b
2623b
3.0b
126.3
21.4a
< 0.05
< 0.001
ns
< 0.05
1624
0.17
28.34
4.57
Lamina (%)
Psuedostems (%)
Dead matter (%)
Reproductive parts (%)
64.7
25.7
9.7
0
58.0
31.3
10.7
0
66.9
25.9
7.2
0
63.3
30.0
6.7
0
ns
ns
ns
10.83
5.64
6.62
Sward surface height (cm)
13.3
13.1
10.9
12.1
ns
3.29
Clover plots
Total pasture DM yield (kg ha-1)
Clover (%)
Weed (%)
1172
95.0
5.0
1148
93.8
6.2
1145
94.5
5.5
1272
94.2
5.8
ns
ns
ns
149.5
9.08
9.08
Sward surface height (cm)
8.3
7.7
7.5
7.5
ns
1.60
Grass plots
61
ME (range 12.8 to 13.5 MJ ME kg-1 DM) was high in all forages with timothy having the
lowest value (Table 4.7). N% was the lowest in both ryegrasses, although both values
remained relatively high (>3.5%). Clover plots had lower WSC and NDF contents than
grass plots (Table 4.7).
Table 4.7 Nutritive value of grass and clover determined by NIRS prior to the September
preference test. P values are from one way ANOVA. Means with same subscript within a
row are not significantly different according to LSD test following a significant ANOVA.
TetRG
DipRG
Tim
CF
Clover
P - value
LSD
ME (MJ ME kg DM)
13.5a
13.2bc
12.8d
13.3b
13.2c
< 0.001
0.10
N (%DM)
3.6d
3.9c
4.8b
4.5c
5.0a
< 0.001
0.13
Water soluble carbohydrate
(%DM)
35.6a
31.7b
20.8d
27.6c
17.0e
< 0.001
1.05
Digestibility (%DM)
89.8a
88.0b
84.2d
87.4b
85.5c
< 0.001
0.65
Neutral detergent fibre (%DM)
23.6c
25.9b
28.3a
22.0d
17.3e
< 0.001
0.90
-1
4.2.2.2 Grazing behaviour and pasture changes
There were no significant differences in grazing behaviour among grass species (Table 4.8).
Sheep spent around 33 to 49 % of grazing time on clover. Selection coefficients were all
below 1 but did not differ with grass treatments. The proportion of DMI for clover was
slightly higher (<12%) in cocksfoot than in the other treatments. There were no significant
effects of grass species on decline in height except timothy declined the quickest in the first
two hours (Figure 4.3a). On average grass height declined 7.6 cm, and clover height
declined 4.1 cm. The percentage of grazed petioles and decline in pasture mass did not
differ among species. (Figure 4.3b, Table 4.8).
62
Table 4.8 Grazing behaviour and changes in pasture for the September preference test.
Means with same subscript within a row are not significantly different according to LSD
test following a significant ANOVA. DMI = dry matter intake. SSH = sward surface
height.
TetRG
DipRG
Tim
CF
P - value
LSD
Grazing behaviour
Total grazing (min)
Grazing grass (min)
Grazing clover (min)
Ruminating (min)
Idling (min)
Grazing time on clover (%)
Selection coefficient for clover
DMI for clover in diet (%)
264
175
89
48
127
33.5
0.62
42.9
290
191
103
49
97
35.5
0.59
39.4
259
135
124
24
156
48.2
0.67
42.1
269
159
109
32
139
40.5
0.58
50.9
ns
ns
ns
ns
ns
ns
ns
ns
46.5
53.1
38.0
20.9
47.8
14.45
0.33
36.8
Pasture changes
Decline of clover SSH (cm, overall)
Decline of grass SSH (cm, overall)
Decline of clover SSH (cm, 1st 2hrs)
Decline of grass SSH (cm, 1st 2hrs)
5.2
8.3
2.1
6.1a
4.1
8.4
1.9
5.4a
3.9
6.2
1.8
2.0b
4.0
7.5
1.4
4.8a
ns
ns
ns
< 0.05
1.37
2.76
0.75
2.67
Grazed petioles (%)
56.0
62.0
62.4
66.2
ns
23.59
Decline of clover mass (%)
52.1
50.9
51.0
55.7
ns
20.31
63
(a)
15
10
Height (cm)
Height (cm)
5
0
12
(b)
10
8
6
Height (cm)
4
Tetraploid perennial ryegrass Trt.
Diploid perennial ryegras s Trt.
Timothy Trt.
Cocks foot Trt.
2
0
9:00
11: 00
13: 00
15: 00
17: 00
Time
Figure 4.3 Sward surface height (cm) of (a) grass and (b) clover plots in paired
comparisons during the September trial. The vertical bars represent the LSD for comparing
the differences among grass treatments at the start and end of grazing occasion.
4.2.3
October preference test
4.2.3.1 Pasture characteristics
Grass pasture mass ranged from 2473 to 2869 kg DM ha-1 but did not differ among
treatments (Table 4.9). Cocksfoot plots contained higher proportion of weeds (Table 4.9).
Both ryegrasses had more tillers per ㎡ than timothy and cocksfoot (Table 4.9), but fewer
green leaves per tiller (Table 4.9). The proportion of lamina was the greatest in tetraploid
64
ryegrass followed by timothy, cocksfoot and diploid ryegrass. The proportion of
psuedostems followed the reverse pattern (Table 4.9). Clover had a lower pasture mass
than all the grass plots, and contained a small percentage of weeds (<15%), mainly Poa
annua.
Table 4.9 Pasture mass and pasture composition of grass and clover plots in each grass
treatment combination prior to the October grazing. P values are from one way ANOVA.
Means with same subscript within a row are not significantly different according to LSD
test following a significant ANOVA.
TetRG
DipRG
Tim
CF
P - value
LSD
2473
88.2
3.1b
8.8
2542
85.9
5.0b
9.2
2869
85.3
1.6b
13.1
2674
82.0
12.1a
5.9
ns
ns
< 0.05
ns
1249.4
8.44
5.64
5.11
No. of tillers per m
No. of greenleaves per tiller
Leaf length (mm)
Sheath tube length (mm)
5980a
2.2c
155.1
28.9
4571a
2.2c
141.2
59.2
3219b
3.1a
165.6
48.6
2693b
2.8b
172.6
53.6
< 0.05
< 0.001
ns
ns
2024
0.26
58.3
28.82
Lamina (%)
Psuedostems (%)
Dead matter (%)
Reproductive parts (%)
62.7a
28.2d
9.1
0
39.0d
51.4a
9.7
0
52.9b
33.8c
13.3
0
49.6c
43.7b
6.7
0
< 0.01
< 0.001
ns
7.68
4.32
5.34
Sward surface height (cm)
16.6
20.6
17.9
18.9
ns
4.35
Clover plots
Total pasture DM yield (kg ha-1 )
Clover (%)
Weed (%)
2070
85.7
14.3
2118
86.2
13.8
1992
90
10.0
1928
92.2
7.8
ns
ns
ns
547
14.94
0.15
Sward surface height (cm)
10.7
11.6
11.6
11.8
ns
2.18
Grass plots
-1
Total pasture DM yield (kg ha )
Grass (%)
Weed (%)
Dead (%)
2
65
ME (range 12.6 to 13.1 MJ ME kg-1 DM) was high in all forages with cocksfoot having the
lowest value (Table 4.10). N% was lowest in both ryegrasses, although both values
remained relatively high (>4.0%). Clover plots had lower WSC and NDF contents than
grass plots. WSC concentration was the greatest in tetraploid ryegrass, followed by diploid
ryegrass, timothy and cocksfoot (Table 4.10).
Table 4.10 Nutritive value of grass and clover by NIRS prior to the October preference test.
P values are from one way ANOVA. Means with same subscript within a row are not
significantly different according to LSD test following a significant ANOVA.
TetRG
DipRG
Tim
CF
Clover
P - value
LSD
ME (MJ ME kg DM)
13.1a
12.7c
13.0ab
12.6d
12.8bc
< 0.01
0.20
N (%DM)
4.2c
4.0c
4.7b
4.6b
5.2a
< 0.001
0.41
Water soluble carbohydrate
(%DM)
24.7a
23.5b
22.1c
17.4d
14.7e
< 0.001
1.14
Digestibility (%DM)
87.2a
84.5bc
85.4ab
83.1c
84.1bc
< 0.01
1.46
Neutral detergent fibre (%DM)
28.3b
31.0a
25.8c
30.8a
17.3d
< 0.001
1.68
-1
Timothy had the highest proportion of N and P among grasses, and the lowest in Ca and Na
(Table 4.11). Clover was higher in N, P and Ca and total macronutrient proportion, and
lower in S than grasses (Table 4.11). Both ryegrasses had a higher S concentration than
cocksfoot and timothy (Table 4.11). Timothy contained the highest concentration of P, and
the lowest concentration of Ca and Na (Table 4.11). Of note is the very low Na
concentration in timothy.
66
Table 4.11 The percentage of macronutrients (%DM) contained in laminae of grass and
clover prior to the October grazing.
Tet RG
Dip RG
Tim
CF
WC
P%
0.32
0.32
0.40
0.33
0.42
K%
3.0
2.6
2.9
2.6
2.9
S%
0.44
0.41
0.35
0.37
0.33
Ca%
0.36
0.39
0.31
0.34
1.06
Mg%
0.16
0.15
0.15
0.14
0.19
Na%
0.32
0.32
< 0.01
0.22
0.16
4.2.3.2 Grazing behaviour and pasture changes
Total grazing time was the greatest in the two ryegrasses, intermediate in cocksfoot and
lowest in timothy (Table 4.12). Grazing time on grass, clover and ruminating time did not
differ with grass species. Idling time was longer in timothy than other grass species. The
percentage of time spent grazing clover ranged from 23% to 36%, but was not affected by
species (Table 4.12). Selection coefficients were all below 1 but did not differ with grass
treatments (Table 4.12). DMI% was lower in timothy than the other treatments, its
percentage was more than 50% lower than tetraploid perennial ryegrass. There were no
significant effects of grass species on decline in height or clover mass. On average grass
height declined 9.9 cm, and clover height declined 4.6 cm. The percentage of grazed
petioles was the highest in tetraploid and diploid perennial ryegrass, intermediate in
cocksfoot, and the lowest in timothy (Table 4.12).
67
Table 4.12 Grazing behaviour and changes in pasture for the October preference test.
Means with same subscript within a row are not significantly different according to LSD
test following a significant ANOVA. DMI = dry matter intake. SSH = sward surface
height.
TetRG
DipRG
Tim
CF
P - value
LSD
Grazing behaviour
Total grazing (min)
Grazing grass (min)
Grazing clover (min)
Ruminating (min)
Idling (min)
Grazing time on clover (%)
Selection coefficient for clover
DMI for clover in diet (%)
228ab
143
84
112
95b
36.4
0.50
42.8
236a
181
56
84
115b
23.0
0.40
33.6
198c
144
54
56
181a
27.9
0.25
18.4
217b
156
62
91
127b
28.0
0.40
31.1
< 0.01
ns
ns
ns
< 0.01
ns
ns
ns
13.7
40.9
47.4
39.9
39.7
19.02
0.49
37.3
Pasture changes
Decline of clover SSH (cm, overall)
Decline of grass SSH (cm, overall)
Decline of clover SSH (cm, 1st 2hrs)
Decline of grass SSH (cm, 1st 2hrs)
Grazed petioles (%)
4.38
8.31
2.1
3.7
53.3a
4.01
10.22
2.3
6.0
47.1ab
4.87
9.74
3.5
6.5
17.9c
5.12
11.13
1.9
6.4
30.6bc
ns
ns
ns
ns
< 0.01
1.976
1.886
2.65
4.37
17.42
36.6
26.4
12.8
30.8
ns
35.86
Decline of clover mass (%)
68
30
(a)
25
20
15
Height (cm)
10
Height (cm)
5
0
16
(b)
14
12
10
8
Height (cm)
6
Tetraploid perennial ryegrass Trt.
Diplooid perennial ryegrass Trt.
Timothy Trt.
Cocksfoot Trt.
4
2
0
9:00
11:00
13:00
15:00
17:00
Time
Figure 4.4 Sward surface height (cm) of (a) grass and (b) clover plots in paired
comparisons during the October trial. The vertical bars represent the LSD for comparing
the differences among grass treatments at the start and end of grazing occasion.
4.2.4
November preference test
4.2.4.1 Pasture characteristics
Grass pasture mass ranged from 1694 to 2203 kg DM ha-1 but did not differ among
treatments (Table 4.13). There was a higher proportion of weeds in timothy and cocksfoot
than tetraploid and diploid ryegrass (Table 4.13). Both ryegrasses had more tillers per ㎡
than timothy and cocksfoot (Table 4.13), but fewer green leaves per tiller (Table 4.13). The
69
proportion of lamina was the greatest in tetraploid ryegrass followed by timothy, cocksfoot
and diploid ryegrass. The proportion of psuedostems followed the reverse pattern (Table
4.13). Clover had a lower pasture mass than all the grass plots, and contained a small
percentage of weed (<10%), mainly Poa annua.
Table 4.13 Pasture mass and pasture composition of grass and clover plots in each grass
treatment combination prior to the November grazing. P values are from one way ANOVA.
Means with same subscript within a row are not significantly different according to LSD
test following a significant ANOVA.
TetRG
DipRG
Tim
CF
P - value
LSD
2203
93.1
1.1c
5.8
1784
94.2
2.0b
3.8
1722
92.2
4.2ab
3.7
1694
92.3
5.7a
2.0
ns
ns
< 0.05
ns
831.4
3.17
3.00
3.70
No. of tillers per m
No. of greenleaves per tiller
Leaf length (mm)
Sheath tube length (mm)
5044a
2.6c
148.3
25.1
4017ab
2.5c
134.8
40.9
2627bc
3.3a
131.5
25.5
1781c
2.9b
142.2
37.1
< 0.05
< 0.001
ns
ns
1591.3
0.27
31.26
16.60
Lamina (%)
Psuedostems (%)
Dead matter (%)
Reproductive parts (%)
63.5a
30.6b
5.8
0
43.9b
52.3a
3.7
0
59.7a
36.4b
3.8
0
45.9b
49.5a
2.1
2.5
< 0.05
< 0.01
ns
ns
11.53
11.12
3.80
2.18
Sward surface height (cm)
16.0
17.1
13.7
14.9
ns
3.81
Clover plots
-1
Total pasture DM yield (kg ha )
Clover (%)
Weed (%)
1820
94.0
6.0
1687
91.0
9.0
1644
90.2
9.8
1769
92.9
7.1
ns
ns
ns
296.0
8.72
8.72
Sward surface height (cm)
10.1
9.6
9.8
10.1
ns
1.39
Grass plots
-1
Total pasture DM yield (kg ha )
Grass (%)
Weed (%)
Dead (%)
2
70
ME (range 12.3 to 13.0 MJ ME kg-1 DM) was high in all forages with timothy and diploid
ryegrass having the lowest value (Table 4.14). N% was lowest in diploid ryegrass, with
other grasses having similar N%. Clover plots had lower WSC and NDF contents than
grass plots (Table 4.14).
Table 4.14 Nutritive value of grass and clover determined by NIRS prior to the November
preference test. P values are from one way ANOVA. Means with same subscript within a
row are not significantly different according to LSD test following a significant ANOVA.
TetRG
DipRG
Tim
CF
Clover
P - value
LSD
12.8ab
12.4cd
12.6bc
12.3d
13.0a
< 0.001
0.22
N (%DM)
4.2b
3.9c
4.6b
4.4b
5.2a
< 0.01
0.55
Water soluble carbohydrate
(%DM)
21.0a
20.5a
19.7a
16.3b
13.8c
< 0.001
1.47
Digestibility (%DM)
84.6a
81.9bc
82.9b
81.0c
85.1a
< 0.01
1.72
Neutral detergent fibre (%DM)
28.7a
31.1a
25.3b
31.1a
15.5c
< 0.001
2.46
ME (MJ ME kg-1 DM)
71
4.2.4.2 Grazing behaviour and pasture changes
There were no significant differences in grazing behaviour among grass species (Table
4.15) for any variable measured. On average grass height declined 9.4 cm, and clover
height declined 5.5 cm. The selection coefficient was higher in tetraploid ryegrass than the
other treatments although did not differ significantly among treatments and all values were
below 1.0 (Table 4.15). DMI% for clover was slightly lower in diploid perennial ryegrass
than in the other treatments.
Table 4.15 Grazing behaviour and changes in pasture for the November preference test.
Means with same subscript within a row are not significantly different according to LSD
test following a significant ANOVA. DMI = dry matter intake. SSH = sward surface
height.
TetRG
DipRG
Tim
CF
P - value
LSD
Grazing behaviour
Total grazing (min)
Grazing grass (min)
Grazing clover (min)
Ruminating (min)
Idling (min)
Grazing time on clover (%)
Selection coefficient for clover
DMI for clover in diet (%)
202
106
96
62
149
48.1
0.71
58.9
231
127
104
76
107
44.9
0.53
50.7
223
124
99
48
142
45.0
0.63
61.4
232
114
118
58
123
52.8
0.54
60.6
ns
ns
ns
ns
ns
ns
ns
ns
39.9
41.4
52.4
22.0
50.4
21.08
0.19
31.8
Pasture changes
Decline of clover SSH (cm, overall)
Decline of grass SSH (cm, overall)
Decline of clover SSH (cm, 1st 2hrs)
Decline of grass SSH (cm, 1st 2hrs)
Grazed petioles (%)
5.5
10.1
3.5
6.3
62.1
5.1
10.0
2.2
4.2
63.1
5.6
8.4
3.2
3.8
61.2
5.7
8.9
2.8
5.3
57.2
ns
ns
ns
ns
ns
1.20
3.06
1.36
5.41
25.12
Decline of clover mass (%)
69.4
57.5
61.7
60.8
ns
26.72
72
(a)
20
15
Height (cm)
Height (cm)
10
5
0
14
(b)
12
10
8
6
Height (cm)
4
Tetraploid perennial ryegrass Trt.
Diploid perennial ryegrass Trt.
Timothy Trt.
Cocksfoot Trt.
2
0
9:00
11:00
13:00
15:00
17:00
Time
Figure 4.5 Sward surface height (cm) of (a) grass and (b) clover plots in paired
comparisons during the November trial. The vertical bars represent the LSD for comparing
the differences among grass treatments at the start and end of grazing occasion.
4.2.5
December preference test
4.2.5.1 Pasture characteristics
Grass pasture mass ranged from 2692 to 3397 kg DM ha-1 but did not differ among
treatments (Table 4.16). There was a higher proportion of weeds in cocksfoot than the
other species (Table 4.16). Both ryegrasses had fewer green leaves per tiller than the other
grasses (Table 4.16). Clover had a lower pasture mass than all the grass plots, and
73
contained a small percentage of weed (<10%), mainly Poa annua. All the grass species
except timothy had reproductive parts in this month but overall a low percentage of tillers
remained reproductive (<15%) (Table 4.16).
Table 4.16 Pasture mass and pasture composition of grass and clover plots in each grass
treatment combination prior to the December grazing. P values are from one way ANOVA.
Means with same subscript within a row are not significantly different according to LSD
test following a significant ANOVA.
TetRG
DipRG
Tim
CF
P - value
LSD
3397
95.9
1.6b
2.5
3124
94.2
1.7b
4.1
2889
92.7
3.1ab
4.2
2692
89.6
7.1a
3.3
ns
ns
< 0.05
ns
821.3
6.0
4.44
3.99
No. of tillers per m
No. of greenleaves per tiller
Leaf length (mm)
Sheath tube length (mm)
4711
2.7b
193
31.2
5116
2.4c
164.5
46.6
3499
3.6a
184.2
35.9
2807
2.8b
189.9
37.0
ns
2431.0
< 0.001
0.31
ns
39.88
ns
13.11
Lamina (%)
Psuedostems (%)
Dead matter (%)
Reproductive parts (%)
44.3
47.8
2.5
5.4
32.5
48.3
4.2
15.0
50.3
45.4
4.3
0
46.9
42.4
3.5
7.2
ns
ns
ns
ns
12.26
5.91
4.13
10.59
Sward surface height (cm)
24.5
25.2
22.5
24.1
ns
3.06
Clover plots
-1
Total pasture DM yield (kg ha )
Clover (%)
Weed (%)
3065
96.8
3.2
2888
93.7
6.3
2968
96.5
3.5
2939
96.1
3.9
ns
ns
ns
288.4
7.13
7.13
Sward surface height (cm)
16.8
16.4
16.5
18.0
ns
1.96
Grass plots
-1
Total pasture DM yield (kg ha )
Grass (%)
Weed (%)
Dead (%)
2
74
Metabolisable energy concentration was moderate in diploid ryegrass, timothy and
cocksfoot, and high in tetraploid ryegrass and clover (Table 4.17). The percentage of N was
high in clover. Water soluble carbohydrate was higher for tetraploid and diploid ryegrass
and the lowest for clover. Neutral detergent fibre was significantly high in diploid ryegrass
and low in clover (Table 4.17).
Table 4.17 Nutritive value of grass and clover determined by NIRS prior to the December
preference test. P values are from one way ANOVA. Means with same subscript within a
row are not significantly different according to LSD test following a significant ANOVA.
TetRG
DipRG
Tim
CF
ME (MJ ME kg DM)
12.2b
11.4d
11.8c
11.8c
12.6a
< 0.001
0.21
N (%DM)
3.1bc
2.6c
3.3b
3.6b
4.9a
< 0.001
0.61
Water soluble carbohydrate
(%DM)
23.0a
21.3a
18.1b
13.0c
14.6c
< 0.001
2.03
Digestibility (%DM)
79.6b
73.5d
75.8c
76.4c
81.6a
< 0.001
1.55
Neutral detergent fibre (%DM)
35.8b
43.3a
32.5c
38.1b
18.1d
< 0.001
3.04
-1
Clover P - value
LSD
Diploid ryegrass had the lowest percentage of N and K (Table 4.18). Cocksfoot contained
the highest N, P, K, S and Ca in the grasses. Of note is the very low Na concentration in
timothy. White clover was had a high concentration of N, P, K, Mg, and Ca, and a low
concentration of Na relative to the grasses (Table 4.18).
75
Table 4.18 The percentage of macronutrients (%DM) contained in laminae of grass and
clover prior to the December grazing.
Tet RG
Dip RG
Tim
CF
WC
P%
0.28
0.27
0.32
0.32
0.36
K%
2.5
2.0
2.5
2.6
2.6
S%
0.34
0.30
0.27
0.35
0.28
Ca%
0.34
0.34
0.41
0.41
1.07
Mg%
0.17
0.16
0.18
0.17
0.24
Na%
0.38
0.26
<0.01
0.34
0.22
4.2.5.2 Grazing behaviour and pasture changes
Total grazing time was greater in tetraploid and diploid ryegrasses than cocksfoot and
timothy. Selection coefficient was the lowest in timothy but all values were below 1.0 and
did not significantly differ across the treatments. DMI% was around 1.3 to 1.5 times higher
in cocksfoot than in the other treatments. Decline in grass height was lower in cocksfoot
than other three grass species (Table 4.19).
76
Table 4.19 Grazing behaviour and changes in pasture for the December preference test.
Means with same subscript within a row are not significantly different according to LSD
test following a significant ANOVA. DMI = dry matter intake. SSH = sward surface
height.
TetRG
DipRG
Tim
CF
P - value
LSD
Grazing behaviour
Total grazing (min)
Grazing grass (min)
Grazing clover (min)
Ruminating (min)
Idling (min)
Grazing time on clover (%)
Selection coefficient for clover
DMI for clover in diet (%)
245b
132
113
81
116bc
45.9
0.42
38.1
281a
113
167
69
92c
59.2
0.43
43.2
216b
99
117
79
146ab
53.8
0.36
37.9
225b
72
153
63
153a
68.9
0.51
56.1
P < 0.05
ns
ns
ns
P < 0.05
ns
ns
ns
39.7
71.0
74.6
24.5
35.1
25.55
0.16
25.8
Pasture changes
Decline of clover SSH (cm, overall)
Decline of grass SSH (cm, overall)
Decline of clover SSH (cm, 1st 2hrs)
Decline of grass SSH (cm, 1st 2hrs)
Grazed petioles (%)
7.1
14.2a
3.0
10.1
47.8
8.3
15.2a
3.6
6.3
49.3
6.3
13.2a
3.2
7.4
43.1
8.9
11.6b
5.0
5.1
43.3
ns
P < 0.05
ns
ns
ns
2.94
2.25
2.70
4.00
25.75
Decline of clover mass (%)
42.4
43.3
37.6
44.8
ns
26.87
77
30
(a)
20
Height (cm)
Height (cm)
10
0
(b)
20
15
10
Height (cm)
Tetr aploid perennial ryegr ass Trt.
Diploid perennial ryegrass Trt.
Timothy Trt.
Cocksfoot Trt.
5
0
9:00
11:00
13:00
15:00
17:00
Time
Figure 4.6 Sward surface height (cm) of (a) grass and (b) clover plots in paired
comparisons during the December trial. The vertical bars represent the LSD for comparing
the differences among grass treatments at the start and end of grazing occasion.
4.2.6
Summary of percentage of grazing time on white clover
The average percentage of grazing time on clover ranged from 23% to 68.9% and did not
differ significantly among treatments during the period (Figure 4.7). After reaching the
lowest preference in October, the mean value of each grazing treatment increased until
December (Figure 4.7). The mean value of the percentage of grazing time on clover
averaged over the treatments was significantly higher in December and lower in October (P
78
< 0.001). The mean percentage of grazing time on clover during the overall period did not
differ significantly among treatments
Grazimg time on clover (%)
100
80
60
40
20
0
MAY
SEP
OCT
NOV
DEC
Date
Tetraploid ryegrass treatment
Diploid ryegrass treatment
Timmothy treatment
Cocksfoot treatment
Figure 4.7 Proportion of grazing time on clover plots at each test from May to December
2007. The vertical bars represent the LSD for comparing the differences among grass
treatments at a given grazing occasion.
A strong positive correlation coefficient was obtained between the percentage of clover
grazing time and NDF% in grass and the (r = 0.70), and between the percentage of clover
grazing time and the ratio of NDF% to N% of grass (r = 0.68). The correlation between the
percentage of clover grazing time and the N% of grass was negative and less strong (Table
4.20).
79
Table 4.20 Correlation between average percentage of grazing time on clover plots and
NDF% in grass species, between clover grazing time% and N% in grass, between the mean
proportion of grazing time on clover and a ratio of NDF to N at each test from May to
December 2007.
MAY
SEP
OCT
NOV
DEC
Mean grazing time on clover (%)
47.0
39.4
28.8
47.7
57.0
Mean grass NDF (%DM)
30.9
25.0
29.0
29.1
37.4
Mean grass N (%DM)
5.0
4.2
4.4
4.3
3.2
NDF / N
5.7
5.5
6.0
6.1
10.5
Correlation coefficient
NDF x grazing time on clover
0.70
N x grazing time on clover
-0.49
NDF / N x grazing time on clover
0.68
4.3 DISCUSSION
4.3.1
Low partial preference
Across the grazing studies there was little evidence of a high partial preference for white
clover. The preference for white clover based on the percentage of grazing time ranged
from 23% to 68.9%. Only on one occasion (December 2007) did the partial preference
approach 70%. This result contrasts with numerous other studies that have noted that the
partial preference for white clover when offered with ryegrass in unlimited foraging
conditions is around 70%. For example, in the review of Rutter (2006), most values of
partial preference were in the range 60-80%, with this result occurring for sheep, cattle and
goats. However, occasionally values of lower partial preference have been recorded.
80
Cosgrove et al. (1996) found that cattle spent 47% of grazing time on white clover in the
early summer and autumn preference tests (December and May). However, in preference
tests conducted at another time of the year (February, summer), cattle spend 60-70% of
grazing time on white clover. In addition, the proportion of DMI for white clover was also
low (average 45%, range 18-60%). This denotes that sheep grazed more grass than white
clover in most treatments and at most grazing occasions, and exhibited little partial
preference for white clover throughout the five preference tests.
4.3.1.1 Preference for grasses
A number of possible explanations for the low partial preference exist. The first possible
explanation is that the low partial preference for clover may be high potential intake rate of
grass relative to clover, based upon stronger preference for grass species in this study than
white clover, excluding during the December test. Sheep have been shown on occasions to
prefer pastures that have the highest intake rate (Illius et al., 1992). For example, Illius et
al. (1992) showed sheep preferred mixed grass-clover turves which had the highest intake
rate. It is likely that the grass pastures had a high potential intake rate compared to the
clover pastures in this study. The young pastures (< 1 year old) were grazed to a low
pasture mass after each preference test so that the resulting grass pastures at the next
preference test were always green and leafy. High grass intake rate may have depressed
any partial preference result. In support of this general concept, a re-analysis of data in
Williams (2006) showed that the partial preference for clover increased as the ratio of
intake rate of clover to grass increased (A.M. Nicol, 2008, pers. comm.). This stresses the
need in future studies to consider how preference may alter as the intake rate potential of
the grass changes relative to the clover (e.g. as might be achieved through changes in tiller
81
density, height reductions, age of pastures and reproductive stage). High intake rates of
grasses can also be presumed through the green leafy nature of the pastures and the high
ME content (Plates 4.4 and 4.5; Tables 4.3, 4.7, 4.10, 4.14 and 4.17).
Plate 4.4 Fully grazed cocksfoot (October 2007).
Plate 4.5 Fully grazed tetraploid perennial ryegrass (September 2007).
82
4.3.1.2 N content
The second possible explanation for the low partial preference may be the high N content
of the grass herbage. Throughout the study the N % of the grass ranged from 2.6% N to
5.3% N, and was often similar to the clover which ranged from 4.9% N to 5.2% N. Several
studies using confined feeding treatments have observed that animals alter diet selection to
balance their protein intake (e.g. Kyriazakis and Oldham, 1993; Villalba and Provenza,
1997) and studies with sheep indicate that they do prefer grasses of high N content. For
example, Edwards et al. (1993) showed preference for N fertilized grass (4.8% N) over
unfertilized (3.4% N) grass, but did not measure effects on partial preference for clover. In
a further study, Cosgrove et al. (2002) showed that sheep showed a 70:30 partial
preference for perennial ryegrass with concentrations of 45 and 32 g N kg-1 DM (4.5% and
3.2% N, respectively). But, in a study where N fertilizer was applied to ryegrass plots in
grass-clover preference tests, there was no difference in partial preference for clover
between N fertilized grass (4.5% N) and unfertilized grass (3.2% N) (Cosgrove et al.,
2002). Cosgrove et al. (2002) concluded that N is not the reason why animals prefer white
clover, and that manipulating the N concentration in grass will not cause the switch in
preference required for animals to easily satisfy their preference from typical mixed
species pastures that are grass dominant and have a low proportion of clover.
4.3.1.3 Not getting complete preference test
The third possible explanation for low partial preference for clover may be the fact that
preference tests were not conducted over the entire day. Sheep show a diurnal pattern of
preference, preferring clover in the morning and grass in the afternoon (Parsons et al.,
1994a; Penning et al., 1995b). In this study, preference tests ran from 9 am to 5 pm. Thus,
83
it is possible that the first grazing bout of clover was missed. For example, doing some
simple calculations for November 2007 where the first part of the day (from sunrise to 9
am), assuming that animals grazed clover for the duration from sunrise to 9 am, is added
into actually measured minutes for grazing clover shows that the percentage of clover in
the diet improves in each species (tetraploid perennial ryegrass: 48% (actual) vs 72%
(simulated); diploid perennial ryegrass: 45% vs 69%; timothy: 45% vs 70%; cocksfoot:
53% vs 72%).
This indicates that missing clover bout may have reduced partial preference although this
takes no account of grass grazing bouts which are typically observed late in the day
(Parsons et al., 1994a; Penning et al., 1995b). Inclusion of these would move partial
preference back closer to 50%. Furthermore, in short term preference tests with turves
(Newman et al., 1994b), more than 90% preference for clover has been observed.
4.3.1.4 Novelty
The fourth possible explanation for low partial preference for clover may be the effect of
the background the sheep were grazing on prior to preference tests. Sheep have been
observed to show preference for novelty, preferring items that they had not grazed
previously. For example, Parsons et al. (1994a) showed sheep from a grass background
initially showed a strong preference for clover (90%) while those from a clover
background showed a strong preference for grass (70%). These effects lasted less than 1
day. However, in the current study sheep grazed on a predominantly grass background of
low clover content prior to the preference tests. Thus, if anything they should have shown a
stronger, not weaker, preference for clover.
84
4.3.1.5 Fibre and N concentration
The fifth possible explanation for low partial preference for clover may be the effect of
fibre in “grass”, or the balance between N and fibre concentration in “pastures”. The
highest mean value of grazing time percentage on clover plots was found in all the
treatments in December, when grass had the lowest N% and the highest NDF% on over the
whole experimental period (Sections 4.2.1, 4.2.2, 4.2.3, 4.2.4 and 4.2.5). A strong
correlation coefficient was obtained between NDF% in grass and the percentage of clover
grazing time (r = 0.70), and the ratio of NDF% to N% of grass also relatively strongly
correlated with the proportion of grazing time on clover (r = 0.68). These imply the
possible involvement of dietary fibre of grass herbage in grazing on white clover. In
general, dietary fibre is known as an important factor for stability of ruminal function in
terms of effective functioning of the microflora (Forbes, 2007). Perhaps, dietary fibre
concentration may be, not a single determinant, one of the complex factors for N ingestion
by ruminants when diet with high N% was offered. For example, there is a notion that diet
with high N% can be consumed by ruminants if sufficient fibre was offered, namely, that
white clover with high-N% can be fully grazed if grass contained enough fibre. Although
Cosgrove et al. (2002) concluded that N concentration in grass is not an independent factor
to alter grazing preference, the results from their tests still afford scope for considering the
possibility of N as an important element to affect grazing preference and the significance
of balance between N and fibre concentration in pastures. Cosgrove et al. (2002) showed
that sheep preferred high N-fertilised ryegrass to low N ryegrass when these grasses were
offered as alternatives, and sheep also exhibited a strong preference for white clover
(high-N) over ryegrass of two treatments with different levels of N fertiliser rates (highand low-N). Unfortunately, NDF% was not reported in that experiment, but the degree of
NDF concentration within a same grass species can be conjectured as generally N%
85
increases and NDF% decreases according to N fertiliser rates increase (Moller et al., 1996;
McKenzie et al., 2003). Hence, the ratio of NDF% to N% was assumedly lower in the
high-N fertilised ryegrass and higher in the nil-N fertilised ryegrass. Therefore, if these
ryegrasses had enough concentration of fibre, the results from Cosgrove et al. (2002) may
support the concept of N-fibre balance, and N concentration remains as a strong factor to
positively affect grazing preference. Alternatively, the positive correlation between fibre
concentration in grass and grazing time percentage on clover might be because of sheep
simply rejected grazing grass with higher fibre content and then selected clover with less
fibre content. This idea can be applied to the December test when a certain level of
flowering (<15% DM) in grass species was found.
4.3.2
No grass species effect on preference
A feature of the results was that grass species and ploidy did not significantly affect partial
preference for white clover. On no occasion in the 5 preference tests conducted was there
any evidence of a treatment effect of grass species. This is perhaps surprising given
previous evidence of difference in preference between grass species (e.g. Edwards et al.,
1993).
4.3.2.1 Timothy
The failure of timothy to have different preference ranking may reflect the low Na content
of the herbage. Throughout the three occasions of analyses, Na concentration in timothy
was 0.02% or less, and this follows the report by Sherrell (1978). Previous studies indicate
that sheep show preference for herbage containing high Na content. Sheep require 0.07%
Na in the herbage for maintenance (Towers and Smith, 1983). Gillespie et al. (2006)
86
showed sheep were strongly attracted to hill pastures of low Na content (< 0.03% Na)
when coarse agricultural salt (NaCl) was applied. Furthermore, Chiy et al. (1998) showed
sheep exhibited a preference for Na fertilized ryegrass even when the unfertilized ryegrass
had relatively high Na%. Apart from Na%, NDF% in timothy (ave. 28.1%) was also lower
than other grasses (ave. 31.0%) in the current research. If dietary fibre was positively
involved in grazing preference in this study, this might be another part of the grounds of
low preference for timothy.
4.3.2.2 Cocksfoot
Previous studies have shown that sheep show a low preference for clover (Edwards et al.,
1993), but low preference for cocksfoot compared to white clover was not observed in this
study. This may reflect the high N content of the cocksfoot herbage compared to previous
dryland studies where cocksfoot is of much lower N content. However, as the study of
Cosgrove et al. (2002) which showed partial preference was not altered by N fertilizing the
grass, N is not a single rationale for sheep to graze pastures
4.3.2.3 Effect of pasture mass and height
Grazing animals have previously been shown to make foraging decisions on the basis of
pasture mass or pasture height. These factors influence the ease of access and prehension
of pastures and may affect animals’ diet selection and grazing time on the plot (Pulido and
Leaver, 2001; Phillips, 2002). In this study, variation existed in pasture mass and height
among grass species prior to grazing, particularly in the May preference test. However, this
did not appear to affect preference, either the percentage of grazing time on clover or
87
clover intake, as neither of these was affected by grass species at any time. Correlation
coefficient between the percentage of grazing time on clover and the pre-grazing SSH of
white clover was 0.44, and the coefficient between the proportion of grazing time on clover
plots and the pre-grazing SSH of grass species was 0.36. Sward surface height might affect
grazing time or herbage intake to some extent as it did in the previous studies (e.g. Pulido
and Leaver, 2001) since clover SSH relative to grass SSH was constantly lower. If SSH
was the primary factor to limit grazing time on clover plots in this study, animals would
have grazed tall clover more than short ones. Grazing behaviour was, in fact, often
observed in the area of white clover with lower sward height through this research, rather
than in the area of clover with higher sward height (Plate 4.6, 4.7, 4.8 and 4.9). This may
be because taller white clover patches were more likely to contain higher N derived from
dung or urine in the previous grazing occasion than shorter ones, and sheep might avoid
ingesting swards with excessive N. Furthermore, the selection coefficient, which contains a
factor of pre-grazing pasture mass as a variable, did not differ significantly among
treatments at any dates (Table 4.5, 4.8, 4.12, 4.15 and 4.19). These results indicate that at
no time did grass or clover height or mass appear limiting of intake in this study (Penning
et al., 1991), and affect grazing preference to large extent. In other studies over long time
periods (Rook et al., 2002), lower preferences for clover have been observed but only
when both clover and grass have become very short and restrictive of intake. Thus animals
switch to alterative species to maintain intake.
88
Plate 4.6 Laxly grazed clover patch after the preference test (October 2007). Bare patch is
a point where a pre-grazing herbage sample was collected.
Plate 4.7 Fully grazed clover with low sward height (October 2007).
89
Plate 4.8 Ungrazed tall clover in the same paddock as Plate 4.7 (October 2007).
90
Plate 4.9 Contrast between fully grazed and ungrazed clover after preference test
(December 2007). Few flowers of white clover were left in the fully grazed area near the
electric fence. Lower sward height of clover in the area can be recognised from the smaller
size of their leaves compared to the size in the central area in this paddock.
4.3.2.4 Effect of WSC on grazing preference
Grazing animals have previously been shown to make foraging decisions on the basis of
the water soluble carbohydrate concentration of the forage. For example, sheep showed
animals prefer forage of higher water-soluble carbohydrate concentration (Ciavarella et al.,
2000). In this study WSC concentration was high in all grass species, often higher in
tetraploid and diploid perennial ryegrass than timothy and cocksfoot (Table 4.3, 4.7, 4.10,
4.14 and 4.17). But, this had no effect on partial preference. Thus, there is little evidence
that increasing WSC of grass would be a method to reduce the partial preference shown for
91
clover, so alleviating grazing pressure on the clover.
4.3.3
Possible nutritive value concept
Possible nutritive value (pNV) is a function of grazing time or DMI and herbage nutritive
value, and a concept to simulate the diet actually ingested by animals during the preference
test day, although the true value ought to be considered with the presence of weeds and the
extent of grazed psuedostems, which would affect nutritive properties of grass species (i.e.
nutritive values differ between in laminae and in psudostems). For example, when mean
ME % of herbage samples of timothy and white clover was 14.0 and 12.0 MJ ME kg-1 DM,
respectively, and grazing time on grass and clover were 50% and 50%, respectively, the
pNV in ME of the timothy treatment will be 13.0 MJ ME kg-1 DM (i.e. timothy: (14.0 x
0.50 = 7.0) + clover (12.0 x 0.50 = 6.0)). The grazing time percentage can also be replaced
into DMI%. The simple calculation indicated that the magnitude of the difference between
maximum and minimum of pNV in every simulated nutritional property via both grazing
time proportion and DMI% for clover was smaller than that of the four grass species prior
to the tests even when variables (i.e. grazing time% on clover or DMI% for clover) widely
ranged. For instance, DMI% for clover ranged between 18% and 43% in the October test
but the differences of pNV were smaller than those of nutritive values among grass species
(ME: 0.5 (grass) vs 0.3 (pNV); N: 0.7 vs 0.4; WSC: 7.3 vs 4.2; NDF: 5.2 vs 3.0). This
infers the possibility that the animals might have aimed at acquiring and balancing
nutritive properties by adjusting grass-clover intake. This notion is partly supported by a
hypothesis by Rutter et al. (2000) that a partial preference results from the ruminants’
balancing N and carbon components in the diet. Scott and Provenza (2000) demonstrated
that lambs had a strong preference for the nutrients lacking in the basal diet. Hence, it may
be plausible that grazing preference was largely influenced by nutritional factors in the
92
current study.
50
(a)
40
NDF%
30
20
10
0
40
(b)
pNV
(%)
NDF%
30
20
10
0
MAY
SEP
OCT
NOV
DEC
Date
Figure 4.8 Comparison of neutral detergent fibre concentration of dry matter between (a)
snipped grasses (tetraploid and diploid perennial ryegrass, timothy and cocksfoot) prior to
grazing, and (b) possible nutritive values (pNVs) (%) calculated from the percentage of
grazing time on grass and clover by treatments at each date. The vertical bars indicate the
difference between the maximum and minimum of (a) mean nutritive values among the
grass species or (b) of mean pNVs among the treatments of the date.
93
5 CHAPTER FIVE
GENERAL DISCUSSION
There was little evidence of partial preference for clover approaching 60-70% in this study
as has generally been reported. Some of the reasons for this have been suggested, including
a high %N of grass relative to other studies, the study not including the immediate post
grazing period before sunset, which is often dominated by a preference for white clover,
and the high intake rate of grass species relative to white clover. No definitive explanation
to the low partial preference emerges. However, the proposed high intake rate of the new
grass pastures may be a plausible explanation that needs further examination.
A clear feature to emerge from this study was the lack of any significant effect of grass
species on the partial preference shown for white clover. Thus, there was no evidence from
this study, at least, to support the hypothesis that altering grass species could be used as a
method to reduce the grazing preference shown for clover, which might paradoxically
result in a higher proportion of clover in the sward and the diet (Parsons et al., 1991;
Edwards et al., 2008). However, some significant differences among grass species in the
grazing preference tests may change the partial preference shown for white clover
according to the type of alternative grass species when N content of pastures decreased.
For example, tiller number and green leaf number per tiller significantly differed between
two ryegrass species, tetraploid and diploid perennial ryegrass, and other grass species,
timothy and cocksfoot, in most seasons (Table 4.2, 4.6, 4.9, 4.13 and 4.16), and timothy
and cocksfoot contained higher percentage of N compounds in their laminae constantly
94
during this research (Table 4.3, 4.7, 4.10, 4.14 and 4.17). When N content of pastures is
low, the difference of N content among pasture species necessarily would become crucial,
and the ease of biting green pastures (i.e. greater tiller number or more green leaves per
tiller) can be one of the positive factors to acquire higher N intake from pastures, and
consequently ruminants may exhibit partial preference for clover and preference for grass
may vary among offered grass species. It would be important to confirm the generality of
the result of this study using grass species with a lower N content and lower intake rate
potential than those currently used.
An alternative (animal rather than plant based) approach to manipulating pasture and diet
composition may be to use variation in partial preference within animals. Nicol et al.
(2008) (see also Edwards et al.(2008)) discussed this concept with the idea of breeding of
more and less selective lines of stock. Nicol et al. (2008) showed variation among animals
within a flock in diet selection and that siblings and mothers exhibited similar behaviour.
This indicates the possibility of including diet selection into breeding programmes and
developing lines of selective and non selective sheep which could be used in manipulate
diet and pasture (Winder et al., 1995).
The mixed species pasture study indicated that pastures of high quality were created in all
treatments. For example, metabolisable energy content ranged 12.3 to 13.1 MJ ME kg-1
DM in November 2007 and clover content occasionally exceeded 85%. Of note is the very
high white clover content of the timothy and cocksfoot pastures. Due to the lack of any
effect of grass species on grazing preference, it appears the high clover content in timothy
and cocksfoot mixtures reflects in part the slow establishment of these grasses, which
95
allowed more and larger clover plants to establish.
One concern often expressed over high white clover content in pastures is the low DM
production of white clover;
white clover monocultures only produce around 75% the DM
of a fully fertilised grass sward (Harris and Hoglund, 1977). However, the lost DM
production in this study was relatively small in mixtures over the first year (approximately
17-23% of total DM reduction compared to diploid and tetraploid ryegrass, Table 3.5) due
to high clover production in timothy- and cocksfoot-white clover plots, and growth of
timothy pastures was sometimes higher than ryegrass pastures in the first spring.
A positive side of the high clover content in timothy and cocksfoot pastures is the
markedly improved livestock production that is often exhibited with high clover content.
For example, daily growth rates of lamb on white clover (321 g day-1) were markedly
higher that other species (average 244 g day-1) (Brown, 1990). Timothy monocultures also
show high performance in dairy production (Johnson and Thomson, 1996). Timothy-white
clover mixtures also resulted in greater live weight gain in goats than do the combinations
of white clover with the other grass species including perennial ryegrass and cocksfoot
(Stevens et al., 1992). Thus, a key point for farming systems with slow establishing grass
species, is the high clover content that occurs in the first year, which enables very high
livestock growth rates, with animals to be finished quicker with lower maintenance costs.
This would give timothy based pastures and advantage over ryegrass based pastures at
least in the first year. If desired, timothy pastures could be drilled with ryegrass pastures at
a later date to improve the DM production of the pastures.
It is quite difficult to compare pasture productivity of the same grass species between
pasture production tests and grazing preference tests throughout this experiment with
96
accuracy as some factors were different between these two tests. For instance, grazing
frequency is different between these two experiments. Six grazing tests were carried out in
the pasture experiment whereas five occasions in the preference test. Grass plots in the
preference test received urea of 130 kg N ha-1 and accompanied clover plots did 40 kg N
ha-1 while mixed swards in the pasture test did no urea. However, the values of pre-grazing
pasture mass can be used to contrast pasture productivity between these two kinds of
swards to some extent since all plots were sown on the same day and trimmed with a
mower to approximately same height after every grazing test in both types of experiments.
Total dry matter yield of grass and clover, and the sum of the yield of both species in
monocultures and mixed pastures are shown in Table 5.1. Clover monocultures produced
more clover mass than mixtures with grass species during the experimental period (Table
5.1). Comparing tetraploid and diploid perennial ryegrass, timothy and cocksfoot grew
more in pure swards than in mixed swards with white clover (Table 5.1). Total yield was
higher in pure sward plots than mixed pastures in all of the treatments. In terms of total
annual herbage production and clover yield, monoculture plots demonstrated the
superiority over mixed swards (Table 5.1). Despite the lower sowing rates for mixture plots
relative to monoculture plots, the fact that mixture plots received no urea and relatively
great impact through more frequent grazing opportunities, both types of ryegrass plots in
mixtures yielded close or more grass than did monocultures (Table 5.1). This may
emphasise the importance of white clover as companion pasture species in a mixed sward
in terms of growth of perennial ryegrass.
97
Table 5.1 Annual dry matter production of grass, clover and the sum of grass and clover in
pure swards from 8 January to 6 December 2007 and mixed swards from 8 January to 14
December 2007 (kg DM ha-1). The number excluded the dry matter values of weeds and
dead material.
Monocultures
Mixtures
Grass yield
Clover yield
Grass and
clover yield
Grass yield
Clover yield
Gras s and
clover yield
TetRG
9836
8736
18572
9562
1310
10872
DipRG
9029
8169
17198
9258
818
10076
Tim
10637
8373
19010
2182
5936
8118
CF
9509
8601
18110
2679
5311
7990
Monocultures of white clover had higher plant density than did grass-clover mixed swards
in this study (Table 3.7) although this may be in part due to higher sowing rates for
monocultures relative to mixtures (Table 3.2 and 4.1). Higher density of white clover can
be thought as a primary factor to increase clover DM production during the first year of
pasture establishment through comparison of clover production between in monocultures
and in mixtures with grass species. Spatially separated monocultures of clover and grass in
the current study indicated the advantages in total DM production and clover content,
which are more likely to give higher performance on animal production, on condition that
sowing rates, grazing frequency and fertiliser application were unequal between the pasture
production test and the grazing preference test.
This study focussed on grazing preference and alternative species as methods to increase
the clover content of pastures. These are a limited set of methods and other pasture based
methods may be suitable to enhance clover content in the first few years after pasture
98
establishment at least. For example, perennial ryegrass may be drilled into clover
monocultures or binary mixtures of white clover and slow-establishing grass such as
timothy and cocksfoot after clover is fully established (Hurst et al., 2000; Peter, 2004).
Once fully established, clover may be sufficiently competitive with perennial ryegrass.
Indeed the ryegrass may benefit from the high N environment created by the high clover
pastures. Alternatively, lower seeding rates of ryegrass may be used. Peter (2004) stated
that suppression of slow-establishing pasture species such as white clover and timothy by
perennial ryegrass may occur even at sowing rate of 5 kg ha-1 for perennial ryegrass. Hurst
et al. (2000) demonstrated that perennial ryegrass seeding rate of 3.5 kg ha-1 was sufficient
to cause the poor sociability of ryegrass with white clover. Spring sowing of
slow-establishing grass species such as timothy and cocksfoot with white clover may also
minimise the disadvantage of these grasses in total DM yields as the growth rate of these
slow-establishing grasses is greater than that of both types of perennial ryegrass (Section
3.2.3). Finally, clover and ryegrass may be drilled in alternative mono-cultural rows. This
may be large strips (e.g. paddock half grass and half clover) or smaller strips of alternating
15 to 45 cm rows. Large monocultures create difficulty in management but do allow
differential application of N fertilizer to the grass. Narrow strips maintain the advantages
of clover monocultures and mixed pastures within a same paddock (as the clover invades
into the grass). Stock are still able to select the pastures of desired content, and N transfer
is still likely to be effective from the clover to the grass (Rutter et al., 2005a; Ibboston,
2006; Edwards et al., 2008).
Future research
This study shows that new pastures of high clover content can be established with a range
of grass species. The study raises specific issues about the role of grazing preference that
99
need further work.
1. Establish the relationship between intake rate of grass (and clover) and grazing
preference for clover. This could be achieved by drilling grass pastures of the same cultivar
at different seeding rate (density) and spacing, and comparing in preference tests with
clover. Alternatively, pastures of different age could be used as an additional method to
create variation in intake rate.
2. Determine the effect of grass N% relative to clover N% on grazing preference for clover,
including the involvement of other nutritive values such as fibre concentration. This could
be achieved by offering N fertilized grass species in combination with clover species.
3. Determine the role of herbage Na content in partial preference for clover. It is
hypothesised that increasing the Na content of timothy by salt (NaCl) application may
increase the preference shown for the grass to levels higher than other grass species.
100
6 CHAPTER SIX
CONCLUSIONS
Tetraploid and diploid perennial ryegrass-white clover pastures out-yielded timothy- and
cocksfoot -white clover mixtures in total DM production (by approximately 2000-2600 kg
DM ha-1 more pasture) during the first 12 months after sowing. The advantage in DM yield
of pastures in both types of ryegrass mixtures depends greatly upon their higher growth
rates in the early establishing phase of swards, compared with timothy- and
cocksfoot-mixtures. While tetraploid and diploid ryegrass plots had high grass growth rates,
clover growth was suppressed markedly compared to timothy and cocksfoot in the early
phase of pasture establishment. This leads to white clover growth exceeding timothy and
cocksfoot growth in the first year. Pastures of such high white clover content are highly
likely to deliver very high levels of stock performance per head.
It is concluded that the rapid growth rate of a pasture species in the early phase of pasture
establishment was a key factor for limiting clover content in a mixed sward in the first year,
and partial preference for white clover might be of less importance in determining pasture
composition.
Across the grazing preference studies, there was little evidence of partial preference for
white clover. Several possible explanations for the low partial preference exist including
(1) the high N content of the grass herbage, (2) incomplete days in preference test, (3)
novelty, (4) the effect of fibre in grass forage, or the balance between N and fibre
concentration in pastures and (5) high intake rate of grass relative to white clover. However,
101
no clear explanation emerged.
Grass species and ploidy did not affect partial preference for white clover. The low
preference for timothy may be due to low Na and low fibre concentrations in the timothy
herbage. Grass or clover height or mass did not limit intake largely in this study and water
soluble carbohydrate had not observed effect on partial preference. A simple calculation of
nutritive values of the possibly ingested pastures showed the possibility that the animals
might have aimed at acquiring and balancing nutritive properties by adjusting grass-clover
intake, but this needs more precise measurement of intake rate to make definitive
conclusions.
102
7 References
Black, A. D. and Lucas, R. J. 2000. Caucasian clover was more productive than white
clover in grass mixtures under drought conditions. Proceedings of the New Zealand
Grassland Association, 62, 183-188.
Black, A. D., Moot, D. J. and Lucas, R. J. 2002. Seedling development and growth of
white clover, caucasian clover and perennial ryegrass grown in field and controlled
environments. Proceedings of the New Zealand Grassland Association, 64,
197-204.
Black, A. D., Moot, D. J. and Lucas, R. J. 2006. Development and growth characteristics of
Caucasian and white clover seedlings, compared with perennial ryegrass. Grass
and Forage Science, 61, 442-453.
Brock, J. L. 2006. Grazing management of white clover in mixed pastures. Proceedings of
the New Zealand Grassland Association, 68, 303-307.
Brown, C. 1990. An integrated herbage system for Southland and South Otago.
Proceedings of the New Zealand Grassland Association, 52, 119-122.
Caradus, J. R., Woodfield, D. R. and Stewart, A. V. 1996. Overview and vision for white
clover. In: D. R. Woodfield, (ed). White clover: New Zealand's competitive edge.
Proceedings of a joint symposium between Agronomy Society of New Zealand and
New Zealand Grassland Association held at Lincoln University, New Zealand
21-22 November, 1995. Lincoln, New Zealand: Agronomy Society of New Zealand
and New Zealand Grassland Association, 1-6.
Champion, R. A., Orr, R. J., Penning, P. D. and Rutter, S. M. 2004. The effect of the spatial
scale of heterogeneity of two herbage species on the grazing behaviour of lactating
sheep. Applied Animal Behaviour Science, 88, 61-76.
Chapman, D. F. and Kenny, S. R. 2005. Alternative feedbase systems for southern
Australia dairy farms. 3. Economic returns from extra dry matter consumption. In:
Proceedings of the XXth International Grassland Congress, Wageningen, The
Netherlands. Wageningen Academic Publishers. p 463.
Charlton, J. F. L. and Stewart, A. V. 2000. Timothy - the plant and its use on New Zealand
farms. Proceedings of the New Zealand Grassland Association, 62, 147-153.
Chiy, P. C., Al-Tulihan, A. A., Hassan, M. H. and Phillips, C. J. C. 1998. Effects of sodium
and potassium fertilisers on the composition of herbage and its acceptability to
dairy cows. Journal of the science of food and agriculture, 76, 289-297.
Ciavarella, T. A., Dove, H., Leury, B. J. and Simpson, R. J. 2000. Diet selection by sheep
grazing Phalaris aquatica L. pastures of differing water-soluble carbohydrate
content. Australian Journal of Agricultural Research, 51, 757-764.
Clark, D. A. and Harris, S. L. 1996. White clover or nitrogen fertiliser for dairying? In: D.
103
R. Woodfield, (ed). White clover: New Zealand's competitive edge. Proceedings of
a joint symposium between Agronomy Society of New Zealand and New Zealand
Grassland Association held at Lincoln University, New Zealand 21-22 November,
1995. Lincoln, New Zealand: Agronomy Society of New Zealand and New Zealand
Grassland Association, 107-114.
Cosgrove, G. P. 2005. Novel grazing management: making better use of white clover. In:
Proceedings of the South Island Dairy Event, Christchurch, New Zealand. Lincoln
University. p 181-190.
Cosgrove, G. P., Anderson, C. B. and Fletcher, R. H. 1997. Species preference influences
on cattle grazing behaviour. Proceedings of the XVIII international grassland
congress, Session 5, 7-8.
Cosgrove, G. P., Anderson, C. B. and Fletcher, R. H. 1996. Do cattle exhibit a preference
for white clover? In: D. R. Woodfield, (ed). White Clover: New Zealand's
Competitive Edge. Lincoln, NZ: Agronomy Society of New Zealand Special
Publication No. 11/ Grassland Research and Practice Series No. 6, 83-86.
Cosgrove, G. P., Anderson, C. B., Parsons, A. J., Brock, J. L. and Tilbrook, J. C. 2002. Can
nitrogen-fertilised ryegrass substitute for white clover? Proceedings of the New
Zealand Grassland Association, 64, 205-209.
Cosgrove, G. P. and Edwards, G. R. 2007. Control of grazing intake. In: P. V. Rattary, I. M.
Brookes and A. M. Nicol, (eds). Pasture and Supplements for Grazing Animals,
New Zealand Society of Animal Production (Inc.), Occasional Publication No. 14.
Hamilton, NZ: New Zealand Society of Animal Production (Inc.)
Cosgrove, G. P., Hyslop, M. G., Anderson, C. B., Litherland, A. J. and Lambert, M. G. 2003.
Integrating novel forage management into sheep farm systems. Proceedings of the
New Zealand Grassland Association, 65, 75-81.
Crush, J. R. 1987. Nitrogen fixation. In: M. J. Baker and W. M. Williams, (eds). White
clover. Wallingford, UK: C.A.B International, 185-202.
Crush, J. R., Woodward, S. L., Eerens, J. P. J. and MacDonald, K. A. 2006. Growth and
milksolids production in pastures of older and more recent ryegrass and white
clover cultivars under dairy grazing. New Zealand Journal of Agricultural Research,
49, 119-135.
Cullen, N. A. 1958. Pasture establishment studies at Invermay Research Station.
Proceedings of the New Zealand Grassland Association, 20, 138-147.
Davies, D. A. 1988. The effects of grass variety and fertilizer nitrogen on white clover
performance in the uplands. In: Grassland facing the energy crisis. Proceedings of
11th General Meeting European Grassland federation. Troia, Portugal. p 329-332.
Dewhurst, R. J., Davies, D. R. and Merry, R. J. 2000. Microbial protein supply from the
rumen. Animal Feed Science and Technology, 85, 1-21.
Edwards, G. R., Lucas, R. J. and Johnson, M. R. 1993. Grazing preference for pasture
104
species by sheep is affected by endophyte and nitrogen fertility. Proceedings of the
New Zealand Grassland Association, 55, 137-141.
Edwards, G. R., Parsons, A. J. and Bryant, R. H. 2008. Manipulating dietary preference to
improve animal performance. Australian Journal of Experimental Agriculture, 48,
773-779.
Edwards, G. R., Parsons, A. J., Penning, P. D. and Krebs, J. R. 1995. Relationship between
vegetation state and bite dimensions of sheep grazing contrasting plant species and
its implications for intake rate and diet selection. Grass and forage science, 50,
378-388.
Ettema, P. J. and Ledgard, S. F. 1992. Getting the best out of white clover. Proceedings of
the Ruakura Farmers' Conference, 44, 72-76.
Forbes, J. M. 2007. Voluntary food intake and diet selection in farm animals; 2nd edition.
Wallingford, UK: CAB International. 453 pp.
Fothergil, M. and Davies, D. A. 1993. White clover contribution to continuously stocked
sheep pastures in association with contrasting perennial ryegrasses. Grass and
Forage Science, 48, 369-379.
Frame, J. and Boyd, A. G. 1986. Effect of cultivar and seed rate of perennial ryegrass and
strategic fertilizer nitrogen on the productivity of grass/white clover swards. Grass
and Forage Science, 41, 359-366.
Gibb, M. J. and Treacher, T. T. 1984. The performance of weaned lambs offered diets
containing different proportions of fresh perennial ryegrass and white clover.
Animal Production, 39, 413-420.
Gillespie, B. J. 2006. Modifying merino grazing and pasture composition in the high
country by salt application. Master of Agricultural Science thesis, Lincoln
University, Canterbury, New Zealand. 106 pp.
Haggar, R. J., Standell, C. J. and Birnie, J. E. 1985. Occurrence, impact and control of
weeds in newly sown leys. In: Weeds. Pests and Diseases of Grassland and Herbage
Legumes. Occasional Symposium No. 18. British Grassland Society. British
Grassland Society. p 11-19.
Harris, S. L., Clark, D. A. and Jansen, E. B. L. 1997. Optimum white clover content for
milk production. Proceedings of the New Zealand Society of Animal Production, 57,
169-171.
Harris, W. and Hoglund, J. H. 1977. Influences of seasonal growth periodicity and
N-fixation on competitive combining abilities of grasses and legumes. In. The
proceedings of the XIII International Grassland Congress, 239-243.
Hoglund, J. H. and Brock, J. L. 1987. Nitrogen fixation in managed grasslands. In: R. W.
Snaydon, (ed). Managed Grasslands: Analytical studies. Amsterdam: Ecosystems of
the World 17B, Elsevier, 187-196.
Hughes, J. G. 1975. What sheep eat on developed and undeveloped high country. Tussock
105
Grasslands and Mountain Lands Institute Review, 31, 20-30.
Hurst, R. G. M., Black, A. D., Lucas, R. J. and Moot, D. J. 2000. Sowing strategies for
slow-establishing pasture species on a North Otago dairy farm. Proceedings of the
New Zealand Grassland Association, 62, 129-135.
Ibboston, D. R. 2006. Novel sowing strategies to improve white clover production in
mixed pastures. Bachelor of Agricultural Science (Hons) thesis, Lincoln University,
Canterbury, New Zealand
Illius, A. W., Clark, D. A. and Hodgson, J. 1992. Discrimination and patch choice by sheep
grazing grass-clover swards. Journal of Animal Ecology, 61, 183-194.
Jamieson, W. S. and Hodgson, J. 1979. The effect of daily herbage allowance and sward
characteristics upon ingestive behaviour and herbage intake of calves under strip
grazing management. Grass and Forage Science, 34, 273-282.
Johnson, R. J. and Thomson, N. A. 1996. Effect of pasture species on milk yield and milk
composition. Proceedings of the New Zealand Grassland Association, 57, 151-156.
Kemp, P. D., Condron, L. M. and Matthew, C. 1999. Pastures and soil fertility. In: J. G. H.
White and J. Hodgson, (eds). New Zealand Pasture and Crop Science. South
Melbourne, Victoria, Australia: Oxford University Press, 67-82.
Kyriazakis, I. and Oldham, J. D. 1993. Diet selection in sheep: the ability of growing
lambs to select a diet that meets their crude protein (nitrogen x 6.25) requirements.
British Journal of Nutrition, 69, 617-629.
Langer, R. H. M. 1990. Pasture plants. In: R. H. M. Langer, (ed). Pastures: their ecology
and management. Auckland, NZ: Oxford University Press, 39-74.
MacDuff, J. H., Jarvis, S. C. and Roberts, D. H. 1990. Nitrates: leaching from grazed
grassland systems. In: Nitrates Agriculture Water International Symposium, Paris
La Defense 7-8 November 1990. p 405-410.
McKenzie, B. A., Kemp, P. D., Moot, D. J., Matthew, C. and Lucas, R. J. 1999.
Environmental effects on plant growth and development. In: J. G. H. White and J.
Hodgson, (eds). New Zealand Pasture and Crop Science. South Melbourne,
Australia: Oxford University Press, 29-44.
McKenzie, F. R., Jacobs, J. L. and Kearney, G. 2003. Long-term effects of multiple
applications of nitrogen fertiliser on grazed dryland perennial ryegrass/white clover
dairy pastures in south-west Victoria. 3. Botanical composition, nutritive
characteristics, mineral content, and nutrient selection. Australian Journal of
Agricultural Research, 54, 477-485.
Merry, R. J., Leemans, D. K. and Davies, D. R. 2002. Improving the efficiency of silage-N
utilisation in the rumen through the use of grasses high in water soluble
carbohydrate content. In. The proceedings of the International Silage Conference
XIII. Scottish Agricultural College (SAC), Auchincruive, Ayr, : The International
Silage Conference XIII, 374–375.
106
Mizuno, K. 1999. Studies on palatability in varieties of orchardgrass (Dactylis glomerata
L.). Relationship between plant characters among varieties and the palatability
measured by Holstein heifers and the analysis from viewpoints of plant breeding.
Sochi Shikenjo Kenkyu Hokoku = Bulletin of the National Grassland Research
Institute, 58, 61-123.
Mizuno, K., Dohi, H., Harada, H., Akiyama, F., Shioya, S. and Fujimoto, F. 1998. Studies
on palatability in varieties of orchardgrass (Dactylis glomerata L.). 4. Relationship
between palatability of varieties and their mineral elements and some chemical
constituents in spring, summer and autumn seasons. Grassland Science, 43,
439-448.
Moller, S., Edwards, N. J., Parker, W. J., Hodgson, J. and Wilson, G. F. 1996. Nitrogen
application to dairy pasture - the effect of rate and timing of spring nitrogen
applications o the concentration of pasture nutrients. Proceedings of the New
Zealand Society of Animal Production, 56, 276-279.
Moloney, S. C. 1993. Selection, management and use of cocksfoot cultivars in North Island
pastoral farming. Proceedings of the New Zealand Grassland Association, 55,
119-125.
Moot, D. J., Scott, W. R., Roy, A. M. and Nicholls, A. C. 2000. Base temperature and
thermal time requirement for germination and emergence of temperate pasture
species. New Zealand Journal of Agricultural Research, 43, 15-25.
Newman, J. A., Parsons, A. J. and Harvey, A. 1992. Not all sheep prefer clover: diet
selection revisited. Journal of Agricultural Science, Cambridge, 119, 275-283.
Newman, J. A., Parsons, A. J. and Penning, P. D. 1994a. A note on the behavioural
strategies used by grazing animals to alter their intake rates. Grass and forage
science, 49, 502-505.
Newman, J. A., Parsons, A. J., Thornley, J. H. M., Penning, P. D. and Krebs, J. R. 1995.
Optimal diet selection by a generalist grazing herbivore. Functional Ecology, 9,
255-268.
Newman, J. A., Penning, P. D., Parsons, A. J., Harvey, A. and Orr, R. J. 1994b. Fasting
affects intake behaviour and diet preference of grazing sheep. Animal Behaviour,
47, 185-193.
Nicol, A. M. 2008. variation in diet preference between sheep. Proceedings of the New
Zealand Society of Animal Production, 68, 57-58.
Orr, R. J., Cook, J. E., Young, K. L., Champion, R. A. and Rutter, S. M. 2005. Intake
characteristics of perennial ryegrass varieties when grazed by yearling beef cattle
under rotational grazing management. Grass and Forage Science, 60, 157-167.
Orr, R. J., Rutter, S. M., Yarrow, N. H., Champion, R. A. and Rook, A. J. 2004. Changes in
ingestive behaviour of yearling dairy heifers due to changes in sward state during
grazing down of rotationally stocked ryegrass or white clover pastures. Applied
107
Animal Behaviour Science, 87, 205–222.
Parsons, A. J., Edwards, G. R., Chapman, D. F. and Carran, R. A. 2006. How far have we
come: 75 years 'in clover'? Proceedings of the New Zealand Grassland Association,
68, 7-13.
Parsons, A. J., Harvey, A. and Woledge, J. 1991. Plant-animal interactions in a
continuously grazed mixture. II. The role of differences in the physiology of plant
growth and of selective grazing in the performance and stability of species in a
mixture. Journal of Applied Ecology, 28, 619-634.
Parsons, A. J., Newman, J. A., Penning, P. D., Harvey, A. and Orr, R. J. 1994a. Diet
preference of sheep: effects of recent diet, physical state and species abundance.
Journal of Animal Ecology, 63, 465-478.
Parsons, A. J., Thornley, J. H. M., Newman, J. A. and Penning, P. D. 1994b. A mechanistic
model of some physical determinants of intake rate and diet selection in a
two-species temperate grassland sward. Functional ecology, 8, 187-204.
Penning, P. D., Newman, J. A., Parsons, A. J., Harvey, A. and Orr, R. J. 1995a. The
preferences of adult sheep and goats grazing ryegrass and white clover. Annales de
zootechnie, 44, supplement 113.
Penning, P. D., Parsons, A. J. and Newman, J. A. 1993. The effects of group size on
grazing time of sheep. Applied animal behaviour science, 37, 101-109.
Penning, P. D., Parsons, A. J. and Orr, R. J. 1991. Patterns of ingestive behaviour of sheep
continuously stocked on monocultures of ryegrass or white clover. Applied animal
behaviour science, 31, 237-250.
Penning, P. D., Parsons, A. J., Orr, R. J., Harvey, A. and Champion, R. A. 1995b. Intake
and behaviour responses by sheep, in different physiological states, when grazing
monocultures of grass or white clover. . Applied animal behaviour science, 45,
63-78.
Peter, H. R. 2004. Timothy and white clover establishment from four autumn sowing dates
with five rates of ryegrass. Bachelor of Agricultural Science (Hours) thesis, Lincoln
University, Lincoln; NZ
Phillips, C. J. C. 2002. Cattle behaviour and welfare (2nd ed.). Oxford, UK: Blackwell
Science Ltd. 264 pp.
Phillips, C. J. C. and Yousseff, Y. M. I. 2003. The effect of previous experience of four
pasture species on the grazing behaviour of ewes and their lambs. Animal Science,
77, 329-333.
Pulido, R. G. and Leaver, J. D. 2001. Quantifying the influence of sward height,
concentrate level and initial milk yield on the milk production and grazing
behaviour of continuously stocked dairy cows. Grass and Forage Science, 56,
58-67.
Ridout, M. S. and Robson, M. J. 1991. Diet composition of sheep grazing grass/white
108
clover swards: a re-evaluation. New Zealand Journal of Agricultural Research, 34,
89-93.
Rook, A. J., Harvey, A., Parsons, A. J., Penning, P. D. and Orr, R. J. 2002. Effect of
long-term changes in relative resource availability on dietary performance of
grazing sheep for perennial ryegrass and white clover. Grass and Forage Science,
57, 54-60.
Rutter, S. M. 2006. Diet preference for grass and legumes in free-ranging domestic sheep
and cattle: Current theory and future application. Applied Animal Behaviour
Science, 97, 17–35.
Rutter, S. M., Cook, J. E., Young, K. L. and Champion, R. A. 2005a. Spatial scale of
heterogeneity affects diet choice but not intake in beef cattle. In: Proceedings of a
Satellite Workshop of the XXth International Grassland Congress, Pastoral Systems
in Marginal Environments, Wageningen, The Netherlands. Wageningen Academic
Publishers
Rutter, S. M., Molle, G., Decandia, M. and Giovanetti, V. 2005b. Diet preference of
lactating Sarda ewes for annual ryegrass and sulla. In: Adaptation and management
of forage legumes -- strategies for improved reliability in mixed swards,
Proceedings of 1st COST 852 workshop. p 191-194.
Rutter, S. M., Orr, R. J. and Rook, A. J. 2000. Dietary preference for grass and white clover
on sheep and cattle: an overview. In: A. J. Rook and P. D. Penning, (eds). Grazing
Management: The principles and practice of grazing, for profit and environmental
gain, within temperate grassland systems. Occasional Symposium No.34 British
Grassland Society, 2000. Cairn Hotel, Horrogate: Antony Rowe Ltd, Reading, UK,
73-78.
Scott, L. L. and Provenza, F. D. 2000. Lambs fed protein or energy imbalanced diets forage
in locations and on foods that rectify imbalances. Applied Animal Behaviour
Science, 68, 293-305.
Sharp, J. M. 2007. Use of novel spatial presentations of plant species to improve legume
abundance. PhD thesis, Imperial College, London, UK
Sherrell, C. G. 1978. A note on sodium concentrations in New Zealand pasture species.
New Zealand Journal of Experimental Agriculture, 6, 189-190.
Skuodiene, R. and Repsiene, R. 2005. The influence of different management systems on
the productivity of timothy/clover swards. In: Integrating efficient grassland
farming and biodiversity: Proceedings of the 13th International Occasional
Symposium of the European Grassland Federation, Tartu, Estonia, 29-31 August
2005, Tartu, Estonia. Tartu, Estonia: Estonian Grassland Society
Soder, K. J., Rook, A. J., Sanderson, M. A. and Goslee, S. C. 2007. Interaction of plant
species diversity on grazing behaviour and performance of livestock grazing
temperate region pastures. Crop Science, 47, 416-425.
109
Stevens, D. R., Baxter, G. S., Casey, M. J., Miller, K. B. and Lucas, R. J. 1992. A
comparison of six grasses for animal production. Proceedings of the New Zealand
Grassland Association, 54, 147-150.
Stevens, D. R., Casey, M. J., Turner, J. D., Baxter, G. S. and Miller, K. B. 1993. Grasslands
Kahu timothy: quality pasture for animal performance. Proceedings of the New
Zealand Grassland Association, 55, 127-132.
Swift, G., McClelland, T. H., Cleland, A. T., Milne, J. A. and Hunter, E. A. 1993. A
comparison of diploid and tetraploid perennial ryegrass and tetraploid
ryegrass/white clover swards under continuous sheep stocking at controlled sward
heights. 1. Sward characteristics. Grass and Forage Science, 48, 279-289.
Swift, G., Vipond, J. E., Fitzsimons, J., McClelland, T. H., Hunter, E. A. and Milne, J. A.
1992. Feeding value for sheep of a tetraploid perennial ryegrass-white clover sward.
In: Grass on the Move: A positive way forward for the grassland farmer, Occasional
symposium No.26, British Grassland Society: Proceedings of the British Grassland
Society winter meeting held at Great Malvern, Worcestershire, 25-26 November
1991, Great Malvern, Worcestershire. Reading, UK: British Grassland Society. Vol.
26. p 161-164.
Taiz, L. and Zeiger, E. 2002. Plant physiology (3rd Ed). Sunderland, MA: Sinauer
Associates, Inc. 259-282 pp.
Thomas, R. J. 1992. The role of legumes in the nitrogen cycle of productive and
sustainable pastures. Grass and Forage Science, 47, 133-142.
Torres-Rodriguez, A., Cosgrove, G. P., Hodgson, J. and Anderson, C. B. 1997. Cattle diet
preference and species selection as influenced by availability. Proceedings of the
New Zealand Society of Animal Production, 57, 197-198.
Towers, N. R. and Smith, G. S. 1983. Sodium. In: N. D. Grace, (ed). The mineral
requirements of grazing ruminants. New Zealand Society of Animal Production.
Occasional Publication No.9: New Zealand Society of Animal Production
Ulyatt, M. J. 1981. The feeding value of herbage: Can it be improved? New Zealand
Agricultural Science, 15, 200-205.
Ulyatt, M. J., Lancashire, J. A. and Jones, W. T. 1977. The nutritive value of legumes.
Proceedings of the New Zealand Grassland Association, 38, 107-118.
Villalba, J. J. and Provenza, F. D. 1997. Preference for flavoured foods by lambs
conditioned with intraruminal administrations of nitrogen. British Journal of
Nutrition, 78, 545-561.
Waghorn, G. C. and Clark, D. A. 2004. Feeding value of pastures for ruminants. New
Zealand Veterinary Journal, 52, 320-331.
Watkin, B. R. 1975. The performance of pasture species in Canterbury. Proceedings of the
New Zealand Grassland Association, 36, 180-190.
Williams, S. J. 2006. Partial preference of sheep differing dietary preferences for clover
110
and grass at varying availabilities. Bachelor of Agricultural Science (Hons) thesis,
Lincoln University, Canterbury, New Zealand
Winder, J. A., Walker, D. A. and Bailey, C. C. 1995. Genetic aspects of diet selection in the
Chihuahuan desert. Journal of Range Management, 48, 549-553.
Woodfield, D. R., Clifford, P. T. P., Cousins, G. R., Ford, J. L., Baird, I. J., Miller, J. E.,
Woodward, S. L. and Caradus, J. R. 2001. Grasslands Kopu II and Crusader: new
generation white clovers. Proceedings of the New Zealand Grassland Association,
63, 103-108.
111
8 Acknowledgements
This study is founded on the contributions of the following people.
First of all, I would like to express my most gratitude to my supervisor, Dr. Grant Edwards,
for his support and guide, which have made me interested in grazing preference and
pasture science during the research. –Thank you for having me involved in such a terrific
theme.
To my associate supervisors, Dr. Derrick Moot and Dr. Alastair Nicol. I owe them for their
support and advice, especially when I was writing this thesis.
To Dr. Racheal Bryant for herbage analysis and assisting observation tests. Throughout the
field experiment, her assists were really helpful in both field and laboratory, and always
eased my burden.
To staff at Field Service Centre, particularly, Mr. Dave Jack and Mr. James McKenzie for
assisting my field work. Without their help, my field experiments could not even start or
finish in no troubles.
To Dr. Bruce McKenzie. His support practically enabled me to commence this research. I
cannot forget his encouraging words in the beginning of this project.
To Yoshitaka Uchida, Thomas Maxwell and Kurt Molloy for observation tests. They
contributed to the field tests even in a hot or chilly day. In particular, Yoshitaka helped me
in many ways for a long time.
112
To Dave Monks and Fiona Sinclair for collecting pasture establishment data prior to my
field experiment.
To colleagues at my postgraduates office for helping and inspiring me.
To teachers at the language school of Lincoln University, Dr. Madelene Yapp, Ms. Stacey
Boniface and Ms. Joan Melvyn for teaching me English before entering the postgraduate
school. Especially, Madelene has kept encouraging me even after the completion of my
English programme.
To Dr. Junichi Sudo and Professor Kazuaki Araki in Japan for letters of recommendation
for entering Lincoln University.
To my friends in Japan, New Zealand and around the world. Their encouragement always
supports me.
To Japanese farmers, waiting patiently for my return. Their voices are always the strongest
drive for me to tackle the science on grazing.
To Dr. Eric Kawabe for introducing the interest and importance of science on soil, pastures
and grazing animals to me first, and encouraging me all the time. －If had not met you, I
would never had come to New Zealand to study agricultural science.
To Dr. and Mrs. Takase, my beloved uncle and aunt. They have supported me a lot in many
ways, for a long time, especially when I went hardships in Japan.
113
Finally, my most heartfelt thanks – more than words can say - to my mother and father for
being my parents. - You constantly support and love me throughout my life.
114